Decision-making method for person account delivery selection of social media platform

Abstract

The invention belongs to the technical field of the Internet, and relates to a decision-making method for delivery selection of a person account on a social media platform. According to the invention, a set of method is set, comprehensive measurement and calculation are carried out by combining various recent data expressions of the account, and the putting decision suggestion result of the evaluated account is fed back by a centesimal score. The method is worthy of noticing that aiming at the evaluation problem of people in different fan orders, the method originally creates a set of preposed grading system calculation method, so that the account can be evaluated in an environment with the same quantity as the account, and objective and fair guidance can be provided for the result of a delivery selection decision.

Description

Technical Field

[0001] The present application belongs to the field of Internet technology and designs a decision-making method for selecting and placing influencer accounts on social media platforms. Background Art

[0002] As the importance of digital marketing in the field of advertising and marketing gradually increases, brands will allocate more marketing budgets to this field. Among them, influencer placement on social media has become a top priority for brands. According to relevant data from the "2023 China Digital Marketing Trends Research" report, influencer marketing (i.e., influencer placement on social media) has become the top business in brand marketing priorities in 2023. At present, social media platforms such as Douyin, Kuaishou, Weibo, Xiaohongshu, and Bilibili have all launched their own influencer marketing service platforms (namely Juliangxingtu, Magnet Engine, Microtask, Dandelion, and Fireworks), in order to facilitate brand owners to understand the details of influencers (creators who publish content on various platforms), the performance of recent works, and the brands / content cases they have cooperated with.

[0003] However, in terms of account selection and decision-making, advertisers generally choose designated agencies to be responsible for the project. The agencies will choose suitable platforms and influencers based on their own experience, company resources, price and other factors. Although the influencer marketing service platforms of the above-mentioned social platforms have various data dimensions for presentation, there is no unified standard for guiding the decision of "selection" or "non-selection", which largely relies on the experience, judgment and accumulation of agency employees. In the existing decision-making logic, the objective explanation is poor. When customers ask why they choose the account "Zhang San" instead of the account "Li Si", it is impossible to judge with quantitative standards; secondly, the standards are not unified. Due to the different characteristics of each social platform and the different user bases, accounts with large fan levels and accounts with small fan levels on the same platform cannot use the same (multiple) standard values ​​for unified measurement and calculation until a decision is made. Accounts with different fan levels need to be evaluated within their own suitable ranges, that is, it can be intuitively understood that the content of an account with 100,000 fans cannot be directly compared with the content of an account with 10 million fans to assess the pros and cons.

[0004] To this end, the present invention has created a new and scientific social media account placement decision-making method to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide a decision-making method for selecting expert accounts for placement on social media platforms, which can also be called an evaluation method.

[0006] Based on the above-mentioned problems of inconsistent standards and poor interpretability in the selection of influencer accounts on social media platforms, the present invention aims to set up a set of methods to comprehensively measure the recent performance of various data of the account and provide feedback on the recommended results of the placement decision of the evaluated account in the form of a percentage score. It is worth noting that in order to evaluate influencers with different levels of fans, this method has created a set of pre-classification system calculation methods, so that accounts can be evaluated in an environment with the same amount of fans, and can provide objective and fair guidance for the results of the placement selection decision.

[0007] The evaluation method of the present invention includes three standard modules: a creativity evaluation module, a fan influence evaluation module and an estimated delivery effect evaluation module; and one additional module: an account influence module.

[0008] The creativity evaluation module is used to measure the production of works by the account during the evaluation period.

[0009] Among them, the fan influence evaluation module is used to measure the influence of the account on fans during the evaluation period.

[0010] Among them, the estimated delivery effect evaluation module is used to evaluate the expected effect of the work after the account is released.

[0011] Among them, the account influence module assists the above three modules in positioning the social influence of the account.

[0012] Specifically, the functions and roles of each module are as follows:

[0013] Among them, the creativity evaluation module is used for:

[0014] The creativity evaluation module is used to measure the production of works by an account during the evaluation period. On the one hand, the overall content volume is considered based on the cumulative amount of published works, and on the other hand, the content production capacity and activity level of an account during the period are considered based on the number of newly published works during the period.

[0015] Among them, the fan influence evaluation module is used to measure the influence of an account on fans during the evaluation period. On the one hand, the number of fans is considered based on the change in fans during the evaluation period. On the other hand, the quality of fan activity and contribution are considered based on the interaction of fans in the account content and the overall fan activity of the account. Simply put, not only do accounts with many fans or a growing number of fans have good fan influence through quantitative evaluation, but the proportion of fans participating in content interaction and their level of activity also need to be evaluated for quality considerations. Only accounts that perform well in both quantity and quality can be identified as having a good overall fan influence performance, avoiding the current practice of increasing ratings by brushing or buying fans on social platforms.

[0016] Among them, the estimated delivery effect evaluation module is used to evaluate the expected effect of the work after the account is released. Through comprehensive calculation of indicators such as the number of views, reposts, comments, and likes of 15 works arranged in reverse order of release time within the past three months of the account, the expected effect level of the work released by the account is obtained. Finally, multiple results are assigned corresponding weights, summed up, and the results are percentage-based.

[0017] Among them, the account influence module is to assist the above three modules in positioning the social influence of the account. This module mainly conducts comprehensive scoring by identifying the resource level corresponding to the number of account fans, the number of platforms covered by the account, and the number of times the account has been listed in the applicant's own list in the past three months.

[0018] Furthermore, the creativity evaluation module: will perform proportional calculations based on the number of historical cumulative works and the number of newly released works within the evaluation period (the number of original works released and the original work rate of the Weibo platform assessment indicators) and the extreme mean of the large interval where the account's fans are located and the extreme mean of the small interval where the account is located, and perform percentage calculations on the two results according to the corresponding weights. (The large interval refers to the six intervals of super head, head, waist, tail, KOC, and amateur, which are roughly divided according to the descending order of the number of fans of all accounts on the platform. The small interval refers to the difference in the number of fans of accounts on different platforms. Under the large interval, different platforms divide the detailed fan volume layers, and the number of small intervals on different platforms is different.)

[0019] Furthermore, the fan influence evaluation module: the fan play rate, fan like rate, fan comment rate and fan forwarding rate will be calculated based on the number of plays, forwardings, comments, likes and total number of fans of 15 works arranged in reverse order of release time in the past three months, and then the work like rate, work comment rate and work forwarding rate will be calculated based on the average value of the forwarding, commenting and likes of these works and the average value of the number of plays. Then, based on the above-mentioned indicators, the fan quality score of the account is calculated in combination according to the specified weight ratio. According to different platforms, different fan volume levels are divided, and the fan quality scores of all accounts in the level where the evaluated account is located are arranged in descending order and divided into four equal parts, and the upper and lower benchmark values ​​are taken. Finally, the fan quality score of the evaluated account is weighted with the upper and lower benchmark values ​​in the corresponding fan volume level of the platform to obtain the final influence evaluation score.

[0020] Furthermore, the module for evaluating the estimated delivery effect will calculate the average value of the number of views, reposts, comments, likes, and collections (if any) of 15 works arranged in reverse order of release time in the past three months with the extreme means of the large and small groups of the account's fans, and then add and summarize them by weight to get the effect score. In addition, the standard deviation and average of the playback volume, forwarding volume, comment volume, like volume, and collection volume (if any) of the above 15 works will be compared to obtain the stability evaluation of the work; the like volume of the work will be compared with the platform baseline (only the playback volume of the Bilibili platform works is compared, the same below), when the like volume (play volume) of the work is greater than or equal to the upper benchmark value, 4 points will be obtained; when the like volume (play volume) of the work is between the upper benchmark value and the middle benchmark value, 3 points will be obtained; when the like volume (play volume) of the work is between the middle benchmark value and the lower benchmark value, 1 point will be obtained; when the like volume (play volume) of the work is less than the lower benchmark value, no points will be obtained, and finally the scores obtained by the work will be added together to obtain the breakthrough influence evaluation value of the account's work. Add the stability evaluation value of the expert's work to the breakthrough influence evaluation value to obtain the AD value (this concept appears in the patent invention that the company has applied for, and is only cited here for calculation). The effect score and the AD value are weighted and summarized according to the specified ratio to obtain the estimated delivery effect score of the evaluated account.

[0021] Furthermore, the account influence module assigns a fixed score based on the account's fan base, the number of platforms covered by the account's operations, and the number of times the account has been included in the unit's own rankings in the past three months. The three scores are weighted and aggregated according to a specified ratio.

[0022] Finally, the quantitative scores of the four modules are added together to obtain the total score of the account placement decision evaluation, which is named Topklout Prediction Index (TPI).

[0023] Among them, the score of the account influence module is additional points (up to 5 points), which does not affect the overall evaluation of the total score of the three modules of account creativity, fan influence and estimated delivery effect (the total score of these three modules is 100 points).

[0024] Specifically, the evaluation method of the present invention comprises the following steps:

[0025] (1) Based on the platform where the account being evaluated is located, determine the large and small interval standards for the platform account's follower level as follows:

[0026] The major ranges of each platform are divided into six to seven levels: super head, head, waist (pan waist), tail, KOC and amateur. The specific major range standards of each platform are shown in the table below.

[0027]

[0028] Based on the large interval and platform account volume, the account follower level of some platforms is divided into small intervals. The specific small interval standards for the two platforms of Douyin and Kuaishou are as follows

[0029] Magnitude Interval Magnitude Interval First level More than 20 million Ninth weight class 4 million-5 million Second level 15 million to 20 million Tenth weight class 3 million-4 million Third weight class 10 million-15 million Eleventh weight class 2 million-3 million Fourth weight class 9 million-10 million 12th weight class 1 million-2 million Fifth weight class 8 million-9 million Thirteenth weight class 500,000-1,000,000 Sixth weight class 7 million-8 million 14th weight class 100,000-500,000 Seventh weight class 6 million-7 million 15th weight class 20,000-100,000 Eighth weight class 5 million-6 million 16th weight class Below 20,000

[0030] The specific cell interval standards of the Weibo platform are as follows

[0031] Magnitude Interval Magnitude Interval First level More than 30 million Tenth weight class 5 million-6 million Second level 20 million-30 million Eleventh weight class 4 million-5 million Third weight class 15 million to 20 million 12th weight class 3 million-4 million Fourth weight class 12 million to 15 million Thirteenth weight class 2 million-3 million Fifth weight class 10 million-12 million 14th weight class 1 million-2 million Sixth weight class 9 million-10 million 15th weight class 500,000-1,000,000 Seventh weight class 8 million-9 million 16th weight class 100,000-500,000 Eighth weight class 7 million-8 million Seventeenth weight class 20,000-100,000 Ninth weight class 6 million-7 million Eighteenth weight class Below 20,000

[0032] The specific cell interval standards of the Bilibili platform are as follows

[0033] Magnitude Interval Magnitude Interval First level More than 2 million Fifth weight class 50,000-100,000 Second level 1 million-2 million Sixth weight class 10,000-50,000 Third weight class 500,000-1,000,000 Seventh weight class 2000-10,000 Fourth weight class 100,000-500,000 Eighth weight class Below 2000

[0034] The standards for small intervals on the Xiaohongshu platform are consistent with those for large intervals. The standards are as follows:

[0035] Magnitude Interval Magnitude Interval First level More than 1 million Fourth weight class 10,000-100,000 Second level 500,000-1,000,000 Fifth weight class 2000-10,000 Third weight class 100,000-500,000 Sixth weight class Below 2000

[0036] The core calculation logic of the present invention is to perform calculations around the settings of large intervals and small intervals to obtain the corresponding standard - the extreme mean. The extreme mean calculation method of a certain indicator is: take the latest 15 works of all accounts in the interval in the past three months (if there are less than 15 works, take all works) and arrange them in descending order. The value in the upper quartile is the extreme mean of this indicator in the interval, and the rules for large and small intervals are the same.

[0037] For example: to calculate the extreme mean of the large and small intervals of the playback volume of an account with 3.2 million followers on the Douyin platform, take the set of the latest 15 playback volumes of all accounts with 2 million to 4 million followers on the Douyin platform in the past three months, and arrange them in descending order. The upper quartile is the extreme mean of the large interval; take the set of the latest 15 playback volumes of all accounts with 3 million to 4 million followers on the Douyin platform in the past three months, and arrange them in descending order. The upper quartile is the extreme mean of the small interval. (The extreme means of other indicators below are determined in this way and will not be repeated here)

[0038] (2) Obtain the number of followers of the account being evaluated on the designated platform, the number of accumulated works of the account, the number of works published in the past 30 days since the date of evaluation (the Weibo platform distinguishes between original and non-original blog posts), and the number of views, likes, reposts, comments, and favorites (if any) of the 15 most recent works within the three months before the date of evaluation (if there are less than 15 works, obtain all works).

[0039] (3) The creativity score (hereinafter referred to as result A) is calculated based on the account's cumulative number of works, the number of works published in the past 30 days from the date of evaluation (hereinafter referred to as the 30-day number of works), and the extreme mean of the cumulative number of works and the extreme mean of the 30-day number of works between the large and small groups of the account. Among them, the three platforms of Douyin, Kuaishou, and Weibo are required to make a comprehensive calculation based on the cumulative number of works and the 30-day number of works, and the Xiaohongshu platform directly calculates the result A based on the 30-day number of works. The Weibo platform makes a comprehensive calculation based on the number of original blog posts of the account in the past 30 days from the date of evaluation (hereinafter referred to as the 30-day original number) and the proportion of original blog posts (the ratio of the number of original blog posts in the past 30 days from the date of evaluation to the total number of blog posts) and the 30-day original number and the proportion of original blog posts between the large and small groups of the account to obtain the creativity score A.

[0040] (4) Take the number of fans of the evaluated account, the number of plays / reads, likes, comments, reposts, and collections (if any) of the works in the past 30 days, and calculate the average number of plays, likes, comments, reposts, and collections (if any) of the works in the past 30 days respectively. Use the average values ​​of the above 4 (5 if there are collections) to calculate the proportion with the number of fans to obtain the fan play rate, fan like rate, fan comment rate, fan repost rate, and fan collection rate (if any); use the average total number of interactions of the works in the past 30 days to calculate the proportion with the total number of fans to obtain the overlap of interactive fans; select the fan play rate, fan like rate, fan comment rate, fan repost rate, and fan collection rate (if any) and the interactive fan impulse to calculate the proportion together to obtain the preliminary result B'. After calculating the B' value of all accounts on the platform where the evaluated account is located, take the extreme mean of the B' value according to the large and small fan intervals of the evaluated account, and calculate the B' value with the extreme mean of the large and small intervals to obtain the fan influence score B.

[0041] (5) Calculate the average number of views / reads, likes, comments, reposts, favorites (if any), coins (if any), barrage (if any) and interaction rate (the ratio between the sum of likes, reposts, comments, favorites, coins, and barrage and the number of views / reads) of the 15 works of the evaluated account in reverse order of release time in the past three months, and summarize the above average values ​​and perform proportional calculation with the AD value results calculated by the patent previously applied by the Company to obtain the estimated delivery result C.

[0042] (6) Qualitative scoring is performed based on the fan base of the evaluated account, the number of platforms covered by operations, and the number of times the account has been listed on our company's platform list in the past three months. For the fan base of Weibo, Xiaohongshu and Bilibili platforms, the super-top influencers are scored 100 points, the top influencers are scored 80 points, the mid-level influencers are scored 70 points, the bottom influencers are scored 60 points, the KOC is scored 40 points, and the amateurs are scored 20 points. For the fan base of Douyin and Kuaishou platforms, the super-top influencers are scored 100 points, the top influencers are scored 90 points, the mid-level influencers are scored 80 points, the mid-level influencers are scored 70 points, the bottom influencers are scored 60 points, the KOC is scored 40 points, and the amateurs are scored 20 points. For the number of platforms covered by operations, 70 people are covered on 1 platform, 80 points are covered on 2-3 platforms, and 100 points are covered on 4 or more platforms. For the number of times the account has been listed, 100 points are given if it is greater than or equal to 3 times, otherwise 30 points are given for each time listed. Finally, the scores of the three parts are calculated in proportion to obtain the influence score D.

[0043] (7) The above A, B, and C scores are calculated in proportion, with a full score of 100 points; D is an additional score with a full score of 5 points. The higher the total score of A, B, C, and D, the better the recommendation for choosing the account when making account placement decisions.

[0044] (8) The scoring criteria calculated according to (7) are as follows:

[0045] The total score is above 90, and cooperation is recommended.

[0046] The total score is between 80 (inclusive) and 90, and we can cooperate.

[0047] The total score is between 60 (inclusive) and 80. Please try it carefully according to your needs.

[0048] If the total score is below 60, cooperation is not recommended.

[0049] The AD value is calculated as follows:

[0050] a) Obtain the number of views, likes, reposts and comments of the work on at least one platform. The number of interactions is calculated based on the number of likes, reposts and comments.

[0051] b) Calculate the standard deviation and average value based on the number of views and interactions of the work, and obtain the average value of views (hereinafter referred to as result A), the standard deviation of views (hereinafter referred to as result B), the average value of interactions (hereinafter referred to as result C), and the standard deviation of interactions (hereinafter referred to as result D).

[0052] c) Perform algorithm calculation on the average playback volume (hereinafter referred to as result A), the standard deviation of playback volume (hereinafter referred to as result B), the average interaction volume (hereinafter referred to as result C), and the standard deviation of interaction volume to obtain the stability evaluation value of the expert's work.

[0053] d) Compare the number of likes of the work with the platform baseline (only the number of plays on the Bilibili platform is compared, the same below). When the number of likes (plays) of the work is greater than or equal to the upper baseline value, 4 points are obtained; when the number of likes (plays) of the work is between the upper baseline value and the middle baseline value, 3 points are obtained; when the number of likes (plays) of the work is between the middle baseline value and the lower baseline value, 1 point is obtained; when the number of likes (plays) of the work is less than the lower baseline value, no points are obtained. Finally, the scores obtained by the work are added together to obtain the breakthrough influence evaluation value of the account's works.

[0054] e) Add the stability evaluation value of the talent's work to the breakthrough influence evaluation value to obtain the AD value. The calculation process of the AD value can refer to the Chinese patent (application number 2022105164252).

[0055] Through the above technical solution of the present invention, an evaluation model is established based on a large amount of actual delivery case data, and the evaluation method is used to evaluate the decision-making before the social account delivery, and the degree of consistency with the final brand delivery goal reaches more than 85%. Thus, the pre-decision evaluation of social account marketing delivery is realized, and the selection is based on basis and standard.

[0056] Another object of the present invention is to provide an evaluation device for social account delivery selection, including: a creativity evaluation module, a fan influence evaluation module, an estimated delivery effect evaluation module and an account influence module.

[0057] Among them, the functions and roles of the creativity evaluation module, fan influence evaluation module, estimated delivery effect evaluation module and account influence module are the same as above.

[0058] Finally, the quantitative scores of the four modules are added together to obtain the social account placement selection decision evaluation score, which is named Topklout Prediction Index (TPI).

[0059] The present invention can provide objective decision-making criteria for social account delivery selection. The higher the TPI score, the higher the recommendation degree of the account is when making delivery selection, and the greater the possibility of obtaining good results after delivery (the effect here refers to the communication effect based on advertiser revenue considerations, that is, the exposure of the brand / product or the conversion effect, that is, the product purchase brought about by the user after the content is seen), thereby providing strong support for advertisers to achieve their marketing goals.

[0060] Specifically, the four component modules of the evaluation method of the present invention can evaluate the content performance of the expert account from different perspectives, whether it has sufficient productivity, influence, coverage and communication effectiveness, and the evaluation method is scientific, open and transparent. The creativity evaluation module measures the productivity level of the account. If the amount of content released is large (the Weibo platform considers original production, if the amount of original content is large and the proportion of original content is high), the score of this module will be high (the theoretical full score is 100 points), which proves that the account has the ability to continuously output; the fan influence evaluation module can truly evaluate the fan activity of the account, and eliminate the situation where the account has a large number of fans but low active fans, resulting in a watered-down basic number of fans to obtain high influence. Therefore, the higher the proportion of fans participating in the interaction, the higher the fan influence score. In the process of account selection decision-making, the situation of falsely high "zombie fans" of the account can be effectively avoided.

[0061] The most important consideration of the present invention is the estimated delivery effect evaluation module. As mentioned in the explanation of the above scheme, not only the average level of the multi-dimensional performance of the account's works in the past period of time is calculated, but it will also be compared and calculated with the extreme means of the large and small areas where the account's fans are located, so as to locate its performance among accounts of the same level, and make the comparison have standards and dynamically float standards, avoiding the absolute value comparison and evaluation of the communication effect between the top accounts and the middle and tail accounts caused by the large difference in the number of fans, so that the evaluated account can obtain an objective evaluation within a suitable comparison range.

[0062] In addition, the account influence evaluation module is an additional score, which provides a supplementary evaluation of the relatively static industry influence of the account. Although super-head accounts, accounts with multiple operating platforms, and accounts that appear on the company's list many times will receive higher scores, the total additional points are controllable after the proportion calculation and do not affect the 100-point basic score of accounts at all levels. It can not only reflect a certain account influence gain, but also will not cause a difference in the starting point of the evaluation of mid- and low-end accounts (if it is included in the 100-point basic score, the mid- and low-end accounts may naturally only be able to score 90 points or even lower as the full score, resulting in different evaluation upper limits).

[0063] The final evaluation index TPI of the present invention is the sum of the scores of the four modules, and the step-by-step result evaluation standard is confirmed according to the total score. A TPI total score below 60 indicates that the overall performance of the account is poor, and it is not recommended to select the account for placement; a TPI total score between 60-80 indicates that the overall performance of the account is average, and it is recommended to choose a trial placement with caution; a TPI total score between 80-90 indicates that the overall performance of the account is good, and cooperation can be selected; a TPI total score above 90 indicates that the overall performance of the account is excellent, and cooperation is recommended. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 , Flow chart of the evaluation method of the present invention

[0065] Figure 2 , the delivery evaluation test results displayed in the software

[0066] Figure 3 , Actual case evaluation test export results

[0067] Figure 4 , Schematic diagram of the flow chart of the evaluation method of the present invention DETAILED DESCRIPTION

[0068] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0069] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0070] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0071] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0072] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more". The names of the messages or information exchanged between multiple parties in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0073] Example 1: Basic evaluation method

[0074] Figure 1This is a flowchart of a decision-making method for selecting an influencer account for placement on a social media platform provided by the disclosed example. This embodiment can be used to provide guidance and assistance to brand owners when they make decisions on the placement of influencer accounts on social platforms. The auxiliary decision result is finally reflected in a numerical value. The numerical value can be calculated by software. The software must be accessed by an operator with the designated authorization of the applicant's company through a computer web terminal. Optionally, the evaluation result (such as the evaluation result calculated by the software) calculated by the computer web terminal can be accessed and applied. Figure 2 ), can be retained for inspection in the form of electronic documents (such as Figure 3 ).

[0075] The first step is to determine the influencer account to be evaluated, which is the pre-selected account identified by the brand owner.

[0076] Optionally, the target scope may be one or more accounts. If there are multiple accounts, these accounts may be from different platforms.

[0077] Among them, according to the social media platform to which the target assessment account belongs, the large and small ranges of the number of followers of the account on the platform involved are determined.

[0078] The second step is to extract the number of followers, the cumulative number of works, the latest 15 content works published within three months from the evaluation time point, and all works published within the last month, and obtain all the communication effect values ​​of the above works in turn. The communication effect values ​​include the number of readings or playbacks, the number of likes, the number of reposts, the number of comments, the number of collections, the number of coins, the number of bullet screens, and the number of works published in the last month. (The Weibo platform needs to distinguish the number of original blog posts, and the Xiaohongshu platform needs to distinguish between graphic notes and video notes)

[0079] The reading volume refers to the number of users who have browsed a certain non-video content on the platform; the playback volume refers to the number of users who have watched a certain video content on the platform; the like volume refers to the number of users who have expressed their like for a certain content and clicked the corresponding logo or performed similar operations on the platform; the forwarding volume refers to the number of users who have reposted or shared a certain content on the platform; the comment volume refers to the number of users who have posted text comments on a certain content on the platform; the collection volume refers to the number of users who have added a certain content to their favorites for quick review and query in the future; the coin count refers specifically to the number of users on the Bilibili platform who support the use of virtual coins for rewards for a certain content; the barrage volume refers specifically to the number of comments displayed on the content screen that are sent in real time on the Bilibili platform when a certain content is played.

[0080] As shown in Formula 1, the evaluation method of the present invention finally presents a numerical TPI score, with a theoretical maximum score of 105 points, which is composed of a creativity evaluation module (theoretical maximum score of 15 points), a fan influence evaluation module (theoretical maximum score of 40 points), an estimated delivery effect evaluation module (theoretical maximum score of 45 points) and an account influence module (theoretical maximum score of 5 points).

[0081] In the third step, the creativity evaluation module is used to assess the content productivity level of the account. As shown in Formula 2-5, the number of works published by the evaluated account in the past month, the cumulative number of works (the number of original works and the proportion of original works are used on the Weibo platform, and the number of works published in the past month is used on the Xiaohongshu platform) and the extreme mean of the cumulative works in the large and small intervals of the fans of the evaluated account, and the extreme mean of the number of works published in the past month (the extreme mean of the number of original works and the extreme mean of the proportion of original works on the Weibo platform, and the extreme mean of the number of works published in the past month on the Xiaohongshu platform) are calculated to obtain the large interval score A, small interval score B, large interval score C, and small interval score D. Then, the four values ​​of A, B, C, and D are proportionally calculated according to Formula 6 to obtain the talent creativity score.

[0082] In the fourth step, the fan influence evaluation module is used by users to assess the real impact of the content produced by the evaluated account on fans, and to consider the active fans behind its content consumption. As shown in Formula 7-Formula 16, the intermediate value is calculated using the number of fans of the influencer, the average number of views, likes, comments, reposts, and collections of all works in the past month. Then, the active fan adjustment coefficient is calculated according to Formula 17-Formula 19, and the active fan adjustment coefficient is multiplied by the number of fans of the evaluated account as shown in Formula 20 to obtain the [Active Fan Quantity] adjusted by this method. Finally, according to the conditions of Formula 21, based on the large and small interval rules of the number of fans of the evaluated account, the fan influence score is calculated.

[0083] The fifth step is to estimate the delivery effect evaluation module by evaluating the broadcast volume, likes, comments, forwarding volume and other communication effects of the latest 15 works (if there are less than 15 works, all works in the past three months) of the evaluated account, and considering the stability of the work performance (i.e., the AD value of the invention patent applied by the company above) to comprehensively estimate the communication effect of the works to be produced by the delivery decision, and display it in percentage. The higher the score, the better the estimated communication effect. As shown in formula 22-formula 31, the median value is calculated by combining the broadcast volume, likes, comments, forwarding volume, collection volume, coin volume, barrage volume, interaction rate, etc. of the latest 15 works (if there are less than 15 works, all works in the past three months) in the past three months, and then the estimated delivery effect score is calculated according to formula 32-formula 35. Among them, the AD value calculation refers to formula 42-49. The upper benchmark value, the middle benchmark value, and the lower benchmark value are fixed values ​​determined according to the different platforms of the evaluated talents, combined with content dissemination and industry experience. The standards for both Douyin and Kuaishou are more than 1 million, 100,000, and 10,000 likes; the standard for Weibo is more than 100,000, 50,000, and 5,000 likes; the standard for Xiaohongshu is more than 100,000, 10,000 likes, and 1,000 likes; the standard for Bilibili is more than 1 million, 500,000, and 100,000 views.

[0084] The sixth step is that the account influence module is a relatively static qualitative evaluation of the macro influence of the account. It is only an additional module and does not affect the overall evaluation of the decision-making direction of the evaluated account based on the creation and production, fan interaction and estimated delivery effect. As shown in Formula 36-Formula 39, the intermediate values ​​are calculated from the three aspects of account fan level, operation coverage, and number of rankings, and then the account influence score is obtained by comprehensive calculation according to the rules shown in Formula 40 and Formula 41.

[0085] Optionally, the order of steps 3 to 6 can be swapped without affecting the respective calculation results.

[0086] Finally, according to Formula 1, the scores of the four modules are weighted and summed in proportion to obtain the account's delivery decision score (i.e., TPI) at the current time point. The TPI score corresponds to the following levels of delivery selection decisions: a total TPI score below 60 indicates that the account's overall performance is poor, and it is not recommended to select the account for delivery; a total TPI score between 60-80 indicates that the account's overall performance is average, and it is recommended to choose a trial delivery with caution; a total TPI score between 80-90 indicates that the account's overall performance is good, and cooperation can be selected; a total TPI score above 90 indicates that the account's overall performance is excellent, and cooperation is recommended.

[0087] Formula 1: TPI score = creativity score * 15% + fan influence score * 40% + estimated delivery effect score * 45% + account influence score * 5%

[0088] Formula 2:

[0089]

[0090] If A is greater than 20 points, take 20 points; if A is less than 20 points, take the actual value

[0091]

[0092] Formula 3:

[0093] If B is greater than 80 points, take 80 points; if B is less than 80 points, take the actual value

[0094] Formula 4:

[0095]

[0096] If C is greater than 20 points, take 20 points; if C is less than 20 points, take the actual value

[0097] Formula 5:

[0098]

[0099] If D is greater than 80 points, take 80 points; if D is less than 80 points, take the actual value

[0100]

[0101] Formula 6:

[0102] Formula 7: Fan playback (or reading rate) = average playback (or reading) volume of works in the past month / number of fans of the account

[0103] Formula 8: Fan Like Rate = Average Likes of Works in the Last Month / Number of Fans of the Account

[0104] Formula 9: Fan comment rate = average number of comments on works in the past month / number of fans of the account

[0105] Formula 10: Fan forwarding rate = average forwarding volume of works in the past month / number of account fans

[0106] Formula 11: Fan collection rate = average number of collections of works in the past month / number of fans of the account

[0107] Formula 12: Work like rate = average number of likes for works in the past month / average number of views for works in the past month Formula 13: Work comment rate = average number of comments for works in the past month / average number of views for works in the past month Formula 14: Work forwarding rate = average number of forwardings for works in the past month / average number of views for works in the past month Formula 15: Playback interaction activity = (total number of views for works in the past month / number of followers of the account) * 50% + ((total number of likes for works in the past month + total number of comments + total number of forwardings) / number of followers of the account) * 50%

[0108] Formula 16: Interactive fans overlap = (total number of likes + total number of comments + total number of reposts for the works in the past month) / number of fans of the account

[0109] Note that if the final result of Formula 7-Formula 16 is greater than 1, then 1 is taken; otherwise, the actual value is taken.

[0110] Formula 17: [Only for Douyin, Kuaishou, and Bilibili platforms]

[0111] Fan quality adjustment coefficient = ((fan play rate * 35% + fan like rate * 15% + fan comment rate * 20% + fan forwarding rate * 30%) + (work like rate * 30% + work comment rate * 30% + work forwarding rate * 40%)) * 50% + play interaction activity * 50%

[0112] Note that if the final result of formula 17 is greater than 1, a random number between 0.97 and 0.99 is taken; otherwise, the actual value is taken.

[0113] Formula 18: [Only for Weibo platform]

[0114] Fan quality adjustment coefficient = (fan reading rate * 10% + fan like rate * 20% + fan comment rate * 30% + fan forwarding rate * 40%) / (work like rate * 20% + work comment rate * 30% + work forwarding rate * 50%)

[0115] Note that if the final result of formula 18 is greater than 0.6, a random number between 0.52 and 0.57 is taken; otherwise, the actual value is taken.

[0116] Formula 19: [Only for Xiaohongshu platform]

[0117] When all the works of the account in the past month are graphic works,

[0118] Fan quality adjustment coefficient (P) = (fan like rate + fan comment rate + fan forwarding rate + fan collection rate) + interactive fan overlap * 0.9

[0119] When the account has both graphic and text works and video works in the past month,

[0120] Fan quality adjustment coefficient (P+V) = ((like rate of fan posts for graphic works + comment rate of fan posts for graphic works + forwarding rate of fan posts for graphic works + collection rate of fan posts for graphic works) + overlap of interactive fans * 0.5) + ((like rate of fan posts for video works + comment rate of fan posts for video works + forwarding rate of fan posts for video works + collection rate of fan posts for video works) + overlap of interactive fans * 0.5)

[0121] Formula 20: Active fans value = fan quality adjustment coefficient * account number of fans

[0122] Formula 21: Fan quality score = ((active fan value / large interval active fan value*large interval weight)*20%+(active fan value / small interval active fan value*small interval weight)*80%)*100

[0123] Among them, the weights of the large and small intervals of the Douyin platform are as follows:

[0124] If the number of followers of the account is greater than or equal to 1 million, the weight of the large interval is 1, and the weight of the small interval is 1.3

[0125] If the number of followers of the account is less than 1 million, the weight of the large interval is 4 and the weight of the small interval is 8

[0126] Among them, the weight of the large interval on the Weibo platform is 4, and the weight of the small interval is 8

[0127] Among them, the weights of large and small intervals on the Kuaishou platform are as follows:

[0128] If the number of followers of the account is greater than or equal to 5 million, the weight of the large interval is 1, and the weight of the small interval is 1.3

[0129] The number of followers of the account is between 1 million (inclusive) and 5 million. The weight of the large range is 1, and the weight of the small range is 1.8

[0130] If the number of followers of an account is less than 1 million, the weight between large and small accounts is 4.

[0131] Among them, the weights of the large and small areas of the Xiaohongshu platform are as follows:

[0132] If the number of followers of the account is greater than or equal to 10,000, the weight of the large interval is 1 and the weight of the small interval is 4

[0133] If the number of followers of an account is less than 10,000, the weight of the large interval is 1 and the weight of the small interval is 100

[0134] Among them, the weights of the large and small intervals of the Bilibili platform are as follows:

[0135] If the number of followers of the account is greater than or equal to 100,000, the weight of the large interval is 1, and the weight of the small interval is 1.2

[0136] The number of followers of the account is between 10,000 (inclusive) and 100,000. The weight of the large range is 1, and the weight of the small range is 4.

[0137] If the number of followers of the account is less than 10,000, the weight of the large interval is 1 and the weight of the small interval is 2

[0138] Formula 22: Estimated play (read) score = ((average play or read volume of 15 works in the past three months / average play or read volume of 15 works in the past three months in the large range) * 0.2 + (average play or read volume of 15 works in the past three months / average play or read volume of 15 works in the past three months in the small range) * 0.8) * 100

[0139] Formula 23: Likes estimated score = ((average number of likes for 15 works in the past three months / average number of likes for 15 works in the past three months in the large interval) * 0.2 + (average number of likes for 15 works in the past three months / average number of likes for 15 works in the past three months in the small interval) * 0.8) * 100

[0140] Formula 24: Comments estimate score = ((average number of comments on 15 works in the past three months / average number of comments on 15 works in the past three months in the large interval) * 0.2 + (average number of comments on 15 works in the past three months / average number of comments on 15 works in the past three months in the small interval) * 0.8) * 100

[0141] Formula 25: Forward estimate score = ((average forwarding volume of 15 works in the past three months / average forwarding extreme value of 15 works in the past three months in the large interval) * 0.2 + (average forwarding volume of 15 works in the past three months / average forwarding extreme value of 15 works in the past three months in the small interval) * 0.8) * 100

[0142] Formula 26: Estimated collection score = ((average collection volume of 15 works in the past three months / average collection value of 15 works in the past three months in the large interval) * 0.2 + (average collection volume of 15 works in the past three months / average collection value of 15 works in the past three months in the small interval) * 0.8) * 100

[0143] Formula 27: Coin-injection prediction score = ((average coin-injection volume of 15 works in the past three months / average coin-injection volume of 15 works in the past three months in the large interval) * 0.2 + (average coin-injection volume of 15 works in the past three months / average coin-injection volume of 15 works in the past three months in the small interval) * 0.8) * 100

[0144] Formula 28: Danmu estimated score = ((average Danmu volume of 15 works in the past three months / average Danmu extreme value of 15 works in the past three months in the large interval) * 0.2 + (average Danmu volume of 15 works in the past three months / average Danmu extreme value of 15 works in the past three months in the small interval) * 0.8) * 100

[0145] Formula 29: [Only for TikTok, Kuaishou, and Weibo platforms]

[0146] Average interaction rate = (average number of likes + comments + forwarding + coin injection + bullet screen for 15 works in the past three months) / average number of views (or average number of readings) for 15 works in the past three months

[0147] Formula 30: [Only for Bilibili platform]

[0148] Average interaction rate = (average number of likes + comments + reposts + coins + bullet screens for 15 works in the past three months) / average number of views for 15 works in the past three months

[0149] Formula 31: Interaction rate estimate score = ((average interaction rate of 15 works in the past three months / average of the average interaction rate of 15 works in the past three months in the large range) * 0.2 + (average interaction rate of 15 works in the past three months / average of the average interaction rate of 15 works in the past three months in the small range) * 0.8) * 100

[0150] Formula 32: [Only for Douyin and Kuaishou platforms]

[0151] Estimated delivery result = (estimated score of playback volume * 30% + estimated score of comment volume * 25% + estimated score of like volume * 15% + estimated score of interaction rate * 30%) * 0.8 + account AD value * 0.2

[0152] Formula 33: [Only for Weibo platform]

[0153] Estimated delivery result = (estimated reading volume * 25% + estimated forwarding volume * 20% + estimated comment volume * 20% + estimated like volume * 25% + estimated interaction rate * 10%) * 0.8 + account AD value * 0.2

[0154] Formula 34: [Only for Xiaohongshu platform]

[0155] Estimated delivery result = (estimated score of comments * 40% + estimated score of likes * 25% + estimated score of favorites * 35%) * 0.8 + account AD value * 0.2

[0156] Formula 35: [Only for Bilibili platform]

[0157] Estimated delivery result = (estimated score of playback volume * 30% + estimated score of forwarding volume * 10% + estimated score of comment volume * 10% + estimated score of like volume * 15% + estimated score of collection volume * 5% + estimated score of coin volume * 5% + estimated score of barrage volume * 5% + estimated score of interaction rate * 20%) * 0.8 + account AD value * 0.2

[0158] Among them, the account AD value is an evaluation result value elaborated in detail in the company's patent application "A Method for Evaluating Digital Marketing Delivery of Influencer Accounts on Social Media Platforms". It is cited here for calculation, and the calculation method and logic are not repeated here.

[0159] Formula 36: [Only applicable to Douyin and Kuaishou platforms]

[0160]

[0161] Formula 37: [Only applicable to Weibo, Xiaohongshu, and Bilibili platforms]

[0162]

[0163] Formula 38:

[0164] Formula 39:

[0165] The list includes the monthly impact of the platform where the assessed account is located, which is published by our company independently.

[0166]

[0167] Power list (3 issues in the past three months), monthly growth list (3 issues in the past three months), quarterly dark horse list (1 issue in the past three months).

[0168] Formula 40: [Only for Weibo, Kuaishou, Xiaohongshu, and Bilibili platforms]

[0169] Account influence score = fan count score * 35% + operation coverage score * 30% + ranking score * 35%

[0170] Formula 41: [Only for TikTok platform]

[0171] Account influence score = fan count score * 30% + operation coverage score * 20% + ranking score * 50%

[0172] Formula 42: AD value = the stability evaluation value of the talent's work + the breakthrough influence evaluation value of the talent's work

[0173] Formula 43: Interaction volume = likes + reposts + comments

[0174]

[0175] Formula 44:

[0176] Formula 45:

[0177]

[0178] Formula 46:

[0179]

[0180] Formula 47:

[0181]

[0182] Formula 48:

[0183]

[0184] Formula 49:

[0185]

[0186] Example 2: Evaluation method test based on actual business

[0187] Figure 2 , Figure 3 This is an example result of the TPI score of a target account placement decision evaluated by this method for a client approached by our company. The client made a placement decision based on the result and verified it.

[0188] The account is one of the candidates in a certain investment plan of the customer. The application software of our company performs calculation according to the logic in Example 1 to obtain Figure 2 TPI result data in. Among them, Figure 2 As shown in the upper-middle quadrilateral radar chart, the scores of the four modules of the evaluated account are: 100 points for creativity module, 86 points for fan quality module, 72 points for estimated delivery effect module and 91 points for account influence module.

[0189] According to the rule of formula 1, weighted summation is performed, and finally the result is as follows Figure 2 The TPI score shown in the upper right part is 86 points, and the result belongs to the range of 80-90. The overall quality is good, and customers are advised to choose to cooperate.

[0190] Finally, based on the evaluation of this delivery decision, the client selected this account in the delivery. After the delivery, the comprehensive evaluation effect of the account met the client's expectations (based on the target audience reach and new product sales diversion, conversion, etc., which are not within the scope of our company and this evaluation method). Therefore, the comparison between the account delivery decision obtained by the client through this method and the actual effect is in line with expectations. After several delivery decision adjustments between our company and multiple clients, the TPI score and decision recommendations finally obtained by this method match the actual execution results of the client by more than 85%, which can meet the needs of guiding delivery selection decisions in daily actual business.

Claims

1. A method for evaluating the selection of expert accounts on social media platforms, characterized in that: It includes 3 standard modules: Creativity Evaluation Module, Fan Influence Evaluation Module and Estimated Delivery Effect Evaluation Module; And 1 additional module: Account Influence Module The creativity evaluation module is used to measure the production of works by the account during the evaluation period. Among them, the fan influence evaluation module is used to measure the influence of the account on fans during the evaluation period. Among them, the estimated delivery effect evaluation module is used to evaluate the expected effect of the work after the account is released. Among them, the account influence module assists the above three modules in positioning the social influence of the account.

2. The evaluation method according to claim 1, characterized in that: The creativity evaluation module considers the overall content volume based on the cumulative amount of published works, and considers the content production capacity and activity within the period based on the number of newly published works within the period. Finally, the two results are assigned corresponding weights, summed up and percentileized.

3. The evaluation method according to claim 1, characterized in that: The fan influence evaluation module is used to measure the influence of an account on its fans during the evaluation period. On the one hand, the number of fans is considered based on the change in fans during the evaluation period. On the other hand, the quality of fan activity and contribution are considered based on the interaction of fans in the account content and the overall fan activity of the account. Finally, the two results are given corresponding weights, summed up, and the results are percentileized.

4. The evaluation method according to claim 1, characterized in that: The estimated delivery effect evaluation module calculates the expected effect level of the account's works through a comprehensive calculation of the number of views, reposts, comments, likes and other indicators of the 15 works arranged in reverse order of release time in the past three months, and finally assigns corresponding weights to multiple results, sums them up and percentage-izes the results.

5. The evaluation method according to claim 1, characterized in that: The account influence module makes a comprehensive score by identifying the resource level corresponding to the account's fan base, the number of platforms covered by the account, and the number of times the account has appeared on the applicant's own list in the past three months. Finally, the three results are assigned corresponding weights, summed up, and the results are percentileized.

6. The evaluation method according to claim 1, characterized in that: (1) Determine the minimum and maximum follower levels for platform accounts based on the platform where the account being evaluated is located. (2) Obtain the number of followers of the account being evaluated on the designated platform, the number of accumulated works of the account, the number of works published in the past 30 days since the date of evaluation, and the number of views, likes, reposts, comments and favorites of the 15 most recent works in the past three months since the date of evaluation. (3) The creativity score is calculated based on the account's cumulative number of works, the number of works published in the past 30 days since the date of evaluation, and the extreme mean of the cumulative number of works and the extreme mean of the number of works in 30 days among the large and small groups to which the account belongs. Among them, the three platforms of Douyin, Kuaishou and Weibo need to make a comprehensive calculation based on the cumulative number of works and the number of works in 30 days. The Xiaohongshu platform directly calculates the result A based on the number of works in 30 days. The Weibo platform calculates the creativity score A based on the number of original blog posts and the proportion of original blog posts in the past 30 days since the date of evaluation, and the number of original blog posts and the proportion of original blog posts in the past 30 days among the large and small groups to which the account belongs. (4) Take the number of followers of the evaluated account, the number of views / plays, likes, comments, reposts, and favorites of the works in the past 30 days, and calculate the average number of views, average number of likes, average number of comments, average number of reposts, and average number of favorites of the works in the past 30 days. Use the above average values ​​to calculate the ratio of the number of followers to obtain the fan play rate, fan like rate, fan comment rate, fan repost rate, and fan favorite rate; The overlap of interactive fans is calculated by calculating the ratio of the average total number of interactions of works in the past 30 days to the total number of fans. The fan play rate, fan like rate, fan comment rate, fan forwarding rate, fan collection rate and interactive fan impact arc are selected for proportional calculation to obtain the preliminary result B'. After calculating the B' value of all accounts on the platform where the evaluated account is located, the extreme mean of the B' value is taken according to the large and small fan intervals of the evaluated account, and the B' value is proportionally calculated with the extreme mean of the large and small intervals to obtain the fan influence score B. (5) Calculate the average number of views / reads, likes, comments, reposts, favorites (if any), coins (if any), bullet comments (if any), and interaction rate of the 15 works of the evaluated account in the past three months, which are arranged in reverse order of release time. Sum up the above average values ​​and calculate the ratio with the AD value result to obtain the estimated delivery result C. (6) Qualitative scoring is performed based on the fan base of the evaluated account, the number of platforms covered by operations, and the number of times the account has been listed on our company's platform list in the past three months. For the fan base of Weibo, Xiaohongshu and Bilibili platforms, the super-top influencers are scored 100 points, the top influencers are scored 80 points, the mid-level influencers are scored 70 points, the bottom influencers are scored 60 points, the KOC is scored 40 points, and the amateurs are scored 20 points. For the fan base of Douyin and Kuaishou platforms, the super-top influencers are scored 100 points, the top influencers are scored 90 points, the mid-level influencers are scored 80 points, the mid-level influencers are scored 70 points, the bottom influencers are scored 60 points, the KOC is scored 40 points, and the amateurs are scored 20 points. For the number of platforms covered by operations, 70 people are covered on one platform, 80 points are covered on 2-3 platforms, and 100 points are covered on 4 or more platforms. For the number of times the account has been listed, 100 points are given if it is greater than or equal to 3 times, otherwise 30 points are given for each time listed. Finally, the scores of the three parts are calculated in proportion to obtain the influence score D. (7) The above A, B, and C scores are calculated in proportion, with a full score of 100 points; D is an additional score, with a full score of 5 points. The higher the total score of A, B, C, and D, the better the recommendation for choosing the account when making account placement decisions. (8) The scoring criteria calculated according to (7) are as follows: The total score is above 90, and cooperation is recommended. The total score is between 80 (inclusive) and 90, and we can cooperate. The total score is between 60 (inclusive) and 80. Please try it carefully according to your needs. If the total score is below 60, cooperation is not recommended.

7. The evaluation method according to claim 6, characterized in that: in, The AD value is calculated as follows: a) Obtain the number of views, likes, reposts and comments of the work on at least one platform. The number of interactions is calculated based on the number of likes, reposts and comments. b) Calculate the standard deviation and average value based on the number of views and interactions of the work, and obtain the average value of views (hereinafter referred to as result A), the standard deviation of views (hereinafter referred to as result B), the average value of interactions (hereinafter referred to as result C), and the standard deviation of interactions (hereinafter referred to as result D). c) Perform algorithm calculation on the average playback volume (hereinafter referred to as result A), the standard deviation of playback volume (hereinafter referred to as result B), the average interaction volume (hereinafter referred to as result C), and the standard deviation of interaction volume to obtain the stability evaluation value of the expert's work. d) Compare the number of likes of the work with the platform baseline (only the number of plays on the Bilibili platform is compared, the same below). When the number of likes (plays) of the work is greater than or equal to the upper baseline value, 4 points are obtained; when the number of likes (plays) of the work is between the upper baseline value and the middle baseline value, 3 points are obtained; when the number of likes (plays) of the work is between the middle baseline value and the lower baseline value, 1 point is obtained; when the number of likes (plays) of the work is less than the lower baseline value, no points are obtained. Finally, the scores obtained by the work are added together to obtain the breakthrough influence evaluation value of the account's works. e) Add the stability evaluation value of the talent's work and the breakthrough influence evaluation value to obtain the AD value.

8. The evaluation method according to claim 6, characterized in that: The major ranges of each platform are divided into six to seven levels: super head, head, waist (pan waist), tail, KOC and amateur. The specific major range standards of each platform are shown in the table below. Based on the large interval and platform account volume, the account follower level of some platforms is divided into small intervals. The specific small interval standards for the two platforms of Douyin and Kuaishou are as follows The specific cell interval standards of the Weibo platform are as follows The specific cell interval standards of the Bilibili platform are as follows The standards for small intervals on the Xiaohongshu platform are consistent with those for large intervals. The standards are as follows: Formula 1: TPI score = creativity score * 15% + fan influence score * 40% + estimated delivery effect score * 45% + account influence score * 5% Formula 2: If A is greater than 20 points, take 20 points; if A is less than 20 points, take the actual value Formula 3: If B is greater than 80 points, take 80 points; if B is less than 80 points, take the actual value formula 4: If C is greater than 20 points, take 20 points; if C is less than 20 points, take the actual value Formula 5: If D is greater than 80 points, take 80 points; if D is less than 80 points, take the actual value Formula 6: Formula 7: Fan playback (or reading rate) = average playback (or reading) volume of works in the past month / number of fans of the account Formula 8: Fan like rate = average number of likes for works in the past month / number of fans of the account Formula 9: Fan comment rate = average number of comments on works in the past month / number of fans of the account Formula 10: Fan forwarding rate = average number of forwardings of works in the past month / number of fans of the account Formula 11: Fan collection rate = average number of collections of works in the past month / number of fans of the account Formula 12: Work like rate = average number of likes for works in the past month / average number of views of works in the past month Formula 13: Work comment rate = average number of comments on works in the past month / average number of views of works in the past month Formula 14: Work forwarding rate = average number of forwardings of works in the past month / average number of views of works in the past month Formula 15: Playback interaction activity = (total number of plays in the past month / number of followers of the account) * 50% + ((total number of likes + total number of comments + total number of reposts in the past month) / number of followers of the account) * 50% Formula 16: Interactive fans overlap = (total number of likes + total number of comments + total number of reposts for the works in the past month) / number of fans of the account Note: If the final result of Formula 7-Formula 16 is greater than 1, then take 1; otherwise take the actual value Formula 17 [only for Douyin, Kuaishou, Bilibili platforms] Fan quality adjustment coefficient = ((fan play rate * 35% + fan like rate * 15% + fan comment rate * 20% + fan forwarding rate * 30%) + (work like rate * 30% + work comment rate * 30% + work forwarding rate * 40%)) * 50% + play interaction activity * 50% Note that if the final result of formula 17 is greater than 1, a random number between 0.97 and 0.99 is taken; otherwise, the actual value is taken. Formula 18 [Only for Weibo platform] Fan quality adjustment coefficient = (fan reading rate * 10% + fan like rate * 20% + fan comment rate * 30% + fan forwarding rate * 40%) / (work like rate * 20% + work comment rate * 30% + work forwarding rate * 50%) Note that if the final result of formula 18 is greater than 0.6, a random number between 0.52 and 0.57 is taken; Otherwise, take the actual value Formula 19 [Only for Xiaohongshu platform] When all the works of the account in the past month are graphic works, Fan quality adjustment coefficient (P) = (fan like rate + fan comment rate + fan forwarding rate + fan collection rate) + interactive fan overlap * 0.9 When the account has both graphic and text works and video works in the past month, Fan quality adjustment coefficient (P+V) = ((like rate of fan posts for graphic works + comment rate of fan posts for graphic works + forwarding rate of fan posts for graphic works + collection rate of fan posts for graphic works) + overlap of interactive fans * 0.5) + ((like rate of fan posts for video works + comment rate of fan posts for video works + forwarding rate of fan posts for video works + collection rate of fan posts for video works) + overlap of interactive fans * 0.5) Formula 20: Active fans value = fans quality adjustment coefficient * number of fans on the account Formula 21 Fan quality score = ((active fans value / large interval active fans value*large interval weight)*20%+(active fans value / small interval active fans value*small interval weight)*80%)*100 Among them, the weights of the large and small intervals of the Douyin platform are as follows: If the number of followers of the account is greater than or equal to 1 million, the weight of the large interval is 1, and the weight of the small interval is 1.3 If the number of followers of the account is less than 1 million, the weight of the large interval is 4 and the weight of the small interval is 8 Among them, the weight of the large interval on the Weibo platform is 4, and the weight of the small interval is 8 Among them, the weights of large and small intervals on the Kuaishou platform are as follows: If the number of followers of the account is greater than or equal to 5 million, the weight of the large interval is 1, and the weight of the small interval is 1.3 The number of followers of the account is between 1 million (inclusive) and 5 million. The weight of the large range is 1, and the weight of the small range is 1.8 If the number of followers of an account is less than 1 million, the weight between large and small accounts is 4. Among them, the weights of the large and small areas of the Xiaohongshu platform are as follows: If the number of followers of the account is greater than or equal to 10,000, the weight of the large interval is 1 and the weight of the small interval is 4 If the number of followers of an account is less than 10,000, the weight of the large interval is 1 and the weight of the small interval is 100 Among them, the weights of the large and small intervals of the Bilibili platform are as follows: If the number of followers of the account is greater than or equal to 100,000, the weight of the large interval is 1, and the weight of the small interval is 1.2 The number of followers of the account is between 10,000 (inclusive) and 100,000. The weight of the large range is 1, and the weight of the small range is 4. If the number of followers of the account is less than 10,000, the weight of the large interval is 1 and the weight of the small interval is 2 Formula 22: estimated play (read) score = ((average play or read volume of 15 works in the past three months / average play or read volume of 15 works in the past three months in the large range) * 0.2 + (average play or read volume of 15 works in the past three months / average play or read volume of 15 works in the past three months in the small range) * 0.8) * 100 Formula 23 Likes Estimated Score = ((Average Likes of 15 Works in the Last Three Months / Average Likes of 15 Works in the Last Three Months in the Large Interval)*0.2+(Average Likes of 15 Works in the Last Three Months / Average Likes of 15 Works in the Last Three Months in the Small Interval)*0.8)*100 Formula 24: Comment estimate score = ((average number of comments on 15 works in the past three months / average value of average comments on 15 works in the past three months in the large interval) * 0.2 + (average number of comments on 15 works in the past three months / average value of average comments on 15 works in the past three months in the small interval) * 0.8) * 100 Formula 25 Forwarding estimate score = ((average forwarding volume of 15 works in the past three months / average forwarding extreme value of 15 works in the past three months in the large interval) * 0.2 + (average forwarding volume of 15 works in the past three months / average forwarding extreme value of 15 works in the past three months in the small interval) * 0.8) * 100 Formula 26 Collection Estimation Score = ((Average collection volume of 15 works in the past three months / Average collection value of 15 works in the past three months in the large interval) * 0.2 + (Average collection volume of 15 works in the past three months / Average collection value of 15 works in the past three months in the small interval) * 0.8) * 100 Formula 27: Estimated coin-input score = ((average coin-input volume of 15 works in the past three months / average coin-input volume of 15 works in the past three months in the large interval) * 0.2 + (average coin-input volume of 15 works in the past three months / average coin-input volume of 15 works in the past three months in the small interval) * 0.8) * 100 Formula 28: estimated barrage score = ((average barrage volume of 15 works in the past three months / average barrage extreme value of 15 works in the past three months in the large range) * 0.2 + (average barrage volume of 15 works in the past three months / average barrage extreme value of 15 works in the past three months in the small range) * 0.8) * 100 Formula 29 [Only for TikTok, Kuaishou, and Weibo platforms] Average interaction rate = (average number of likes + comments + forwarding + coin injection + bullet screen for 15 works in the past three months) / average number of views (or average number of readings) for 15 works in the past three months Formula 30 [Only for Bilibili platform] Average interaction rate = (average number of likes + comments + reposts + coins + bullet screens for 15 works in the past three months) / average number of views for 15 works in the past three months Formula 31: Estimated interaction rate score = ((average interaction rate of 15 works in the past three months / average interaction rate of 15 works in the past three months in the large interval) * 0.2 + (average interaction rate of 15 works in the past three months / average interaction rate of 15 works in the past three months in the small interval) * 0.8) * 100 Formula 32 [Only for Douyin and Kuaishou platforms] Estimated delivery result = (estimated score of playback volume * 30% + estimated score of comment volume * 25% + estimated score of like volume * 15% + estimated score of interaction rate * 30%) * 0.8 + account AD value * 0.2 Formula 33 [Only for Weibo platform] Estimated delivery result = (estimated reading volume * 25% + estimated forwarding volume * 20% + estimated comment volume * 20% + estimated like volume * 25% + estimated interaction rate * 10%) * 0.8 + account AD value * 0.2 Formula 34 [Only for Xiaohongshu platform] Estimated delivery result = (estimated score of comments * 40% + estimated score of likes * 25% + estimated score of favorites * 35%) * 0.8 + account AD value * 0.2 Formula 35 [Only for Bilibili platform] Estimated delivery result = (estimated score of playback volume * 30% + estimated score of forwarding volume * 10% + estimated score of comment volume * 10% + estimated score of like volume * 15% + estimated score of collection volume * 5% + estimated score of coin volume * 5% + estimated score of barrage volume * 5% + estimated score of interaction rate * 20%) * 0.8 + account AD value * 0.2 The account AD value is an evaluation result value described in detail in the patent application of the company "A method for evaluating digital marketing delivery of social media platform influencer accounts". It is cited here for calculation and the calculation method and logic will not be repeated. Formula 36 [Only applicable to Douyin and Kuaishou platforms] Formula 37 [Only applicable to Weibo, Xiaohongshu, and Bilibili platforms] Formula 38 Formula 39 The list includes the monthly influence of the platform where the evaluated account is located, which is published by our company independently. List (3 issues in the past three months), Monthly Growth List (3 issues in the past three months), Quarterly Dark Horse List (1 issue in the past three months), Formula 40 [Only for Weibo, Kuaishou, Xiaohongshu, and Bilibili platforms] Account influence score = fan level score * 35% + operation coverage score * 30% + ranking score * 35% Formula 41 [Only for Douyin platform] Account influence score = fan volume score * 30% + operation coverage score * 20% + ranking score * 50%, Formula 42: AD value = expert work stability evaluation value + expert work breakthrough influence evaluation value Formula 43: Interaction volume = likes + reposts + comments Formula 44: Formula 45: Formula 46: Formula 47: Formula 48: Formula 49:

9. An evaluation device for selecting expert accounts for placement on a social media platform, characterized in that: It includes 3 standard modules: Creativity Evaluation Module, Fan Influence Evaluation Module and Estimated Delivery Effect Evaluation Module; And 1 additional module: Account Influence Module, The creativity evaluation module is used to measure the production of works by the account during the evaluation period. Among them, the fan influence evaluation module is used to measure the influence of the account on fans during the evaluation period. Among them, the estimated delivery effect evaluation module is used to evaluate the expected effect of the work after the account is released. Among them, the account influence module assists the above three modules in positioning the social influence of the account.

10. The device according to claim 9, characterized in that Creativity evaluation module: The overall content volume is measured by the cumulative amount of published works, and the content production capacity and activity within the period are measured by the number of newly published works within the period. Finally, the two results are given corresponding weights, summed up, and the results are percentage-based; Fan influence evaluation module: used to measure the influence of an account on fans during the evaluation period. On the one hand, the number of fans is considered based on the change in fans during the evaluation period. On the other hand, the quality of fans' activity and contribution are considered based on the fans' interaction in the account content and the overall fan activity of the account. Estimated delivery effect evaluation module: used to evaluate the expected effect of the works released by the account. Through comprehensive calculation of the number of views, forwarding, comments, likes and other indicators of the 15 works of the account arranged in reverse order of release time in the past three months, the expected effect level of the works released by the account is obtained. Finally, multiple results are assigned corresponding weights, summed and the results are percentage-based. Account influence module: It assists the above three modules in positioning the social influence of the account. This module mainly makes a comprehensive score by identifying the resource level corresponding to the number of account fans, the number of platforms covered by the account, and the number of times the account has appeared on the applicant's own list in the past three months.

Images