Information processing device, program, and information processing method

Abstract

An information processing device, a program, and an information processing method for classifying evaluators are provided. [Solution] In the system 10, the information processing device 100 includes an acquisition unit that acquires evaluation information including score information indicating the score of the company 300 that is the evaluation target and comment information indicating comments related to the reason for the score of the evaluation target, and a classification unit that classifies the evaluator who evaluated the evaluation target based on the evaluation information.The classification unit classifies the evaluator as a recommender of the evaluation target when the score of the evaluation target indicated by the score information is higher than a predetermined first score threshold, classifies the evaluator as a neutral person of the evaluation target when the score of the evaluation target is lower than the first score threshold and higher than a predetermined second score threshold that is lower than the first score threshold, and classifies the evaluator as a critic of the evaluation target when the score of the evaluation target is lower than the second score threshold.

Description

[Technical Field]

[0001] The present invention relates to an information processing device, a program, and an information processing method. [Background technology]

[0002] Patent Document 1 describes a cancellation factor analysis system and customer loyalty analysis system that can obtain more accurate analysis results regarding the reasons for canceling content reception contracts. [Prior art document] [Patent Documents] [Patent Document 1] JP 2024-152190 A Summary of the Invention [Means for solving the problem]

[0003] According to one embodiment of the present invention, there is provided an information processing device. The information processing device may include an acquisition unit that acquires evaluation information including score information indicating a score of an evaluation object and comment information indicating a comment related to a reason for the score of the evaluation object. The information processing device may also include a classification unit that classifies evaluators who evaluated the evaluation object based on the evaluation information.

[0004] In the information processing device, the classification unit may classify the evaluator as a recommender of the evaluation target when the score of the evaluation target indicated by the score information is higher than a predetermined first score threshold, may classify the evaluator as a neutral of the evaluation target when the score of the evaluation target is lower than the first score threshold and higher than a predetermined second score threshold that is a score lower than the first score threshold, and may classify the evaluator as a critic of the evaluation target when the score of the evaluation target is lower than the second score threshold.

[0005] In any of the information processing devices, the classification unit may reclassify the evaluators who have been classified as recommenders of the object to be evaluated, neutral people for the object to be evaluated, or critics of the object to be evaluated into recommenders of the object to be evaluated, neutral people for the object to be evaluated, critics of the object to be evaluated, or indifferent people who are indifferent to the object to be evaluated, based on the classification results in which the comments indicated by the comment information are classified into positive comments about the object to be evaluated, neutral comments about the object to be evaluated, negative comments about the object to be evaluated, or comments that are indifferent to the object to be evaluated.

[0006] Any of the information processing devices may further include a calculation unit that calculates the difference between a recommender ratio, which is the ratio of the number of evaluators that the classification unit reclassified into recommenders of the evaluation object to the total number of evaluators who evaluated the evaluation object, and a detractor ratio, which is the ratio of the number of evaluators that the classification unit reclassified into detractors of the evaluation object to the total number of evaluators who evaluated the evaluation object.

[0007] Any of the information processing devices may further include an identification unit that identifies the reason for the score of the evaluation object based on the comment information, and the calculation unit may calculate the difference between the promoter ratio and the critic ratio for each reason for the score of the evaluation object.

[0008] In any of the information processing devices, the acquisition unit may further acquire a prompt for classifying the comment, and the classification unit may classify the comment by inputting the prompt to a generation AI and acquiring from the generation AI the classification result that classifies the comment based on the prompt.

[0009] Any of the information processing devices may further include an information storage unit that stores cancellation rate information indicating the cancellation rate at which the evaluator cancels a contract with the company within a predetermined period from the time the evaluator evaluates the company being evaluated, for each combination of the evaluator category and the comment category, and the classification unit may reclassify the evaluator based further on the cancellation rate information stored in the information storage unit.

[0010] In any of the information processing devices, the acquisition unit may acquire the evaluation information that further includes option information indicating options that correspond to reasons for the score of the evaluation object, and when the classification unit reclassifies the evaluator as an indifferent person who is indifferent to the evaluation object, the classification unit may further classify the evaluator who has been reclassified as an indifferent person who is indifferent to the evaluation object based on the option information.

[0011] Any of the information processing devices may further include an information storage unit that stores the reason for the score of the evaluation object identified from the comment information and cancellation rate information indicating the cancellation rate at which the evaluator will cancel a contract with the company within a predetermined period from when the evaluator evaluated the company that is the evaluation object, for each category of options, and an identification unit that identifies the reason for the score of the evaluation object based on the comment information, and the classification unit may further classify the evaluator who has been reclassified as an indifferent person who is indifferent to the evaluation object based on the cancellation rate information stored in the information storage unit.

[0012] Any of the information processing devices may further include a learning data storage unit that stores learning data including comment data indicating comments related to the reason for the score of the evaluation target and comment category data indicating the category of the comments, and a model generation unit that generates, by machine learning, a decision model from the comment data using the multiple learning data stored in the learning data storage unit as training data, which decides the category of a comment related to the reason for the score of the evaluation target indicated by the comment data, and the classification unit may use the decision model to classify the comment by determining the category of the comment indicated by the comment information from the comment information.

[0013] In any of the information processing devices, the classification unit may reclassify the evaluator, who was classified by the score of the evaluation target indicated by the score information, based on the classification results obtained by classifying the comments indicated by the comment information in accordance with predetermined classification criteria.

[0014] Any of the information processing devices may further include a judgment unit that judges, based on the comment information, whether the reason for the score of the store of the company being evaluated, as indicated by the score information, is attributable to at least one of the store and the store's crew, and the classification unit may classify the evaluator based on the evaluation information when the judgment unit determines that the reason for the score is attributable to at least one of the store and the crew.

[0015] According to one embodiment of the present invention, a program is provided for causing a computer to execute an acquisition procedure for acquiring evaluation information including score information indicating the score of an evaluation object and comment information indicating comments related to the reason for the score of the evaluation object, and a classification procedure for classifying the evaluators who evaluated the evaluation object based on the evaluation information.

[0016] According to one embodiment of the present invention, there is provided an information processing method executed by a computer. The information processing method may include an acquisition step of acquiring rating information including score information indicating a score of an object to be rated and comment information indicating a comment related to a reason for the score of the object to be rated. The information processing method may include a classification step of classifying the evaluators who rated the object to be rated based on the rating information.

[0017] The above summary of the invention does not list all of the necessary features of the present invention, and subcombinations of these features may also constitute inventions. [Brief explanation of the drawings]

[0018] [Figure 1] An example of a system 10 is shown schematically. [Figure 2] FIG. 10 is an explanatory diagram for explaining an example of an evaluation index. [Figure 3] FIG. 10 is an explanatory diagram for explaining an overview of a method for calculating an evaluation index. [Figure 4] FIG. 10 is an explanatory diagram for explaining an example of comparing evaluation targets based on evaluation indices. [Figure 5] 10 shows a schematic diagram of an example of a questionnaire. [Figure 6] 10 shows a schematic diagram of an example of a prompt. [Figure 7] 10A and 10B show schematic examples of labels assigned to comments. [Figure 8] 10 shows an example of the churn rate of the evaluator 250. [Figure 9] FIG. 10 is an explanatory diagram for explaining an example of a method for calculating an evaluation index. [Figure 10] Another example of the churn rate of the evaluator 250 is shown schematically. [Figure 11] FIG. 10 is an explanatory diagram for explaining another example of a method for calculating an evaluation index. [Figure 12] Another example of the system 10 is shown schematically. [Figure 13] FIG. 10 is an explanatory diagram for explaining another example of a method for calculating an evaluation index. [Figure 14] 1 shows an example of a functional configuration of an information processing device 100. [Figure 15] FIG. 2 is an explanatory diagram for explaining an example of a processing flow of the system 10. [Figure 16] An example of the hardware configuration of a computer 1200 that functions as the information processing device 100 is shown in schematic form. DETAILED DESCRIPTION OF THE INVENTION

[0019] In recent years, customer loyalty, which expresses a customer's attachment to and trust in a company and its products and services, has been attracting increasing attention. One reason for this increased attention is that customer loyalty has a stronger correlation with a company's revenue than customer satisfaction, which expresses the satisfaction of customers who have received products and services from a company. Therefore, many companies now use NPS (registered trademark) (Net Promoter Score), a key performance indicator (KPI) that quantifies customer loyalty. NPS is an evaluation metric that quantifies customer loyalty based on the results of categorizing respondents based on their answers to questions asking about customer loyalty. NPS's advantages include being a unified, global evaluation metric and being simple. However, its challenges include the fact that it is calculated based on classification results that may not accurately categorize respondents, and the difficulty of intuitively understanding the effects of NPS-based improvement measures or comparing competitors based on NPS. By solving the issues with NPS, customer loyalty can be quantified with high reliability, and an evaluation index that makes it easy to intuitively understand the effects of competitive comparisons and improvement measures can be realized, so there is a high demand for solving the issues with NPS. The system according to this embodiment employs, for example, a mechanism for classifying evaluators who evaluated the evaluation target based on the scores they gave the evaluation target and comments related to the reasons for their scores. By classifying evaluators based not only on their scores but also on comments related to the reasons for their scores, more appropriate classification of evaluators is possible. In particular, the system according to this embodiment employs, for example, C-NPS (Comment-based NPS), an evaluation index that quantifies customer loyalty from the classification results obtained by classifying customers based on the responses of corporate customers to a question asking about customer loyalty, which are one of a multi-level score scales, and comments related to the reasons for their scores.To calculate C-NPS, customers who respond to questions about customer loyalty can be classified into one of three categories: Promoter, Passive, Detractor, or Indifferent. By calculating C-NPS from the results of appropriately classifying customers based not only on their scores but also on comments related to the reasons for their scores, an evaluation index can be created that can quantify customer loyalty with a high degree of reliability. Furthermore, by calculating C-NPS for each reason for the score based on comments related to the reasons for the score, an evaluation index can be created that makes it easy to compare with competitors and understand the effectiveness of improvement measures.

[0020] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention. In the drawings, the same reference numerals are used to designate the same or similar parts, and redundant explanations may be omitted.

[0021] 1 schematically illustrates an example of a system 10. The system 10 may include an information processing device 100. The system 10 may further include a collection device 400.

[0022] The system 10 may provide, for example, an evaluation service for evaluating an evaluation target. The system 10 may provide, for example, an evaluation support service for supporting the evaluation of an evaluation target. The system 10 may also provide any other service related to the evaluation of an evaluation target.

[0023] The evaluation target is, for example, company 300. The evaluation target is, for example, a store of company 300. The evaluation target is, for example, a service provided by company 300. The evaluation target is, for example, a product of company 300. Note that services and products may be collectively referred to as goods. The evaluation target is, for example, an application of company 300. The evaluation target is, for example, a web page of company 300. Applications and web pages may be included in goods. The evaluation target may be anything else related to company 300.

[0024] 1 shows an example in which the evaluation target is a company 300. The company 300 may be a private company or a public company.

[0025] The evaluator 250 of the evaluation target is, for example, a customer of the company 300. The evaluator 250 of the evaluation target may also be an employee of the company 300. Here, an example in which the evaluator 250 of the evaluation target is a customer of the company 300 will be mainly described.

[0026] A customer of company 300 is, for example, a user who purchases a service provided by company 300. Examples of services include communication services, electronic payment services, customer services, insurance services, medical services, nursing care services, training services, transportation services, and shuttle services. A customer of company 300 may also be a user who purchases a product of company 300. Note that a customer of company 300 may be an ordinary consumer or a corporation.

[0027] The information processing device 100 executes various information processes. For example, the information processing device 100 executes various information processes in order to evaluate an evaluation target.

[0028] The information processing device 100 evaluates the evaluation target by, for example, conducting a questionnaire survey. A questionnaire survey is a survey method in which questions are asked to multiple respondents during a predetermined survey period, and the respondents' answers to the questions are used as data. Examples of the survey period include one day, one week, one month, etc. In the example shown in FIG. 1, the respondents to the questions may be evaluators 250.

[0029] The information processing device 100 evaluates one evaluation target, for example, by conducting a questionnaire survey on the one evaluation target. The information processing device 100 evaluates multiple evaluation targets, for example, by conducting a questionnaire survey on multiple evaluation targets. The multiple evaluation targets are, for example, multiple companies. In this case, the multiple companies may be competing companies. The multiple evaluation targets are, for example, multiple stores of a single company. The multiple evaluation targets may also be multiple products of one or multiple companies. In this case, the multiple products may be similar products.

[0030] The information processing device 100 conducts a questionnaire survey by, for example, sending questions to the communication terminal 200 owned by the evaluator 250. The information processing device 100 sends the questions to the communication terminal 200 via the network 20, for example.

[0031] The network 20 may include a core network provided by a telecommunications carrier. The core network may conform to, for example, a 5G (5th Generation) communication system. The core network may conform to a 6G (6th Generation) communication system or later mobile communication system. The core network may conform to a 3G (3rd Generation) communication system. The core network may conform to an LTE (Long Term Evolution) communication system. The network 20 may include the Internet.

[0032] The information processing device 100 uses, for example, Short Message Service (SMS) to send the question to the communication terminal 200. The information processing device 100 may also send the question to the communication terminal 200 using e-mail.

[0033] The communication terminal 200 may be any communication terminal that can communicate with the information processing device 100. For example, the communication terminal 200 may be a mobile phone such as a smartphone, a tablet terminal, a wearable terminal, etc. The communication terminal 200 may also be a PC (Personal Computer).

[0034] The question may be, for example, a multiple-choice question. In this case, the person answering the question answers the multiple-choice question by selecting an option. The question may be, for example, a comment-type question. In this case, the person answering the question answers the comment-type question by creating a comment.

[0035] The questions include, for example, questions asking about the score of the evaluation object, and questions asking about the reason for the score of the evaluation object.

[0036] The information processing device 100, for example, acquires the answer of the respondent to a question and evaluates the evaluation target based on the acquired answer. The information processing device 100, for example, acquires the answer of the respondent to the question by receiving the answer of the evaluator 250 to the question from the communication terminal 200. The information processing device 100, for example, receives the answer of the evaluator 250 to the question from the communication terminal 200 via the network 20.

[0037] The information processing device 100 acquires evaluation information, for example, as an answer from an answerer to a question. The evaluation information includes, for example, score information indicating the score of the evaluation target. The evaluation information includes, for example, comment information indicating comments related to the reason for the score of the evaluation target. The evaluation information includes, for example, option information indicating options corresponding to the reason for the score of the evaluation target.

[0038] The information processing device 100 classifies the evaluators 250 based on, for example, the evaluation information. The information processing device 100 classifies the evaluators 250 based on, for example, the score of the evaluation target indicated by score information included in the evaluation information. The information processing device 100 classifies the evaluators 250 based on, for example, comments related to the reason for the score of the evaluation target indicated by comment information included in the evaluation information. The information processing device 100 classifies the evaluators 250 based on, for example, options corresponding to the reason for the score of the evaluation target indicated by option information included in the evaluation information.

[0039] The information processing device 100 evaluates the evaluation target based on the classification result obtained by classifying the evaluators 250. The information processing device 100 evaluates the evaluation target, for example, by calculating an evaluation index that quantifies the evaluation of the evaluation target. The information processing device 100 evaluates the evaluation target, for example, by calculating an evaluation index that quantifies the customer loyalty of the evaluation target.

[0040] The collection device 400 collects answers from respondents to questions. The collection device 400 collects answers from respondents to questions by, for example, conducting a questionnaire survey. The collection device 400 collects answers from respondents to questions by, for example, conducting a questionnaire survey on the Web.

[0041] The collection device 400 conducts a questionnaire survey, for example, by transmitting questions to the communication terminal 200 via the network 20. The collection device 400 transmits the questions to the communication terminal 200, for example, by using an application installed on the collection device 400.

[0042] The collection device 400 collects answers from respondents to questions by, for example, collecting evaluation information. The collection device 400 collects evaluation information by, for example, receiving the evaluation information from the communication terminal 200 via the network 20.

[0043] The collection device 400 transmits answers from respondents to the collected questions to the information processing device 100, for example, via the network 20. The collection device 400 transmits collected evaluation information as answers from respondents to the collected questions to the information processing device 100, for example, via the network 20. The information processing device 100 may acquire answers from respondents to the questions collected by the collection device 400 from the collection device 400, and evaluate the evaluation target based on the acquired answers.

[0044] 2 is an explanatory diagram for explaining an example of an evaluation index, which will be mainly described as an example of an evaluation index that quantifies the customer loyalty of the evaluation target.

[0045] The upper diagram in Figure 2 is an explanatory diagram for explaining an example of a conventional evaluation index. Here, we will mainly explain NPS, which is an example of a conventional evaluation index.

[0046] NPS may be an evaluation index that satisfies the relationship "NPS = (promoter rate) - (detractor rate)". The promoter rate is the ratio of the number of promoters who recommend the evaluation target to the total number of evaluators who evaluated the evaluation target. The detractor rate is the ratio of the number of detractors who criticize the evaluation target to the total number of evaluators who evaluated the evaluation target.

[0047] To calculate NPS, respondents are classified as promoters, passives, or detractors based on their response to a customer loyalty question, which is one of a multi-point scale. For example, if the customer loyalty question has an 11-point scale ranging from 0 to 10, respondents who respond with a score of 9 to 10 to the customer loyalty question would be classified as promoters, respondents who respond with a score of 7 to 8 to the customer loyalty question would be classified as passives, and respondents who respond with a score of 0 to 6 to the customer loyalty question would be classified as detractors.

[0048] Detractors may be further classified as either weak detractors or strong detractors. For example, if the customer loyalty question is rated on an 11-point scale from 0 to 10, respondents who answered the customer loyalty question with a score of 5 to 6 would be classified as weak detractors, and respondents who answered the customer loyalty question with a score of 0 to 4 would be classified as strong detractors.

[0049] Examples of questions asking about customer loyalty include, "How likely are you to recommend Company X to a close friend or colleague?" or "On a scale of 0 to 10, how likely are you to recommend Product A from Company Y to an acquaintance?" Questions asking for a rating of an item, including questions asking about customer loyalty, may be examples of multiple-choice questions.

[0050] Promoters are more likely to make positive comments and behave in a positive way towards the target of evaluation to those around them, and are less likely to switch from the target of evaluation to a competitor of the target of evaluation. Neutrals are less likely to make both positive and negative comments and behave in a positive way towards the target of evaluation to those around them, and are more likely than promoters to switch from the target of evaluation to a competitor of the target of evaluation. Detractors are more likely to make negative comments and behave in a negative way towards the target of evaluation to those around them, and are more likely than neutrals to switch from the target of evaluation to a competitor of the target of evaluation.

[0051] In the example shown in the top diagram of Figure 2, in response to a question asking about customer loyalty, A% of respondents were classified as promoters, B% were classified as passives, C% were classified as weak detractors, and D% were classified as strong detractors. Therefore, in the example shown in the top diagram of Figure 2, NPS = A - (C + D).

[0052] The lower diagram in Figure 2 is an explanatory diagram for explaining the differences between the conventional evaluation indexes and the new evaluation indexes. Here, we will mainly explain the differences between the NPS, which is an example of a conventional evaluation index, and the new evaluation indexes.

[0053] One of the challenges of the NPS is that it is an evaluation index calculated based on classification results that may not properly classify respondents. One reason for this inability to properly classify respondents is that response scores tend to be concentrated in the middle. For example, when the response scale is 11 points from 0 to 10, responses to questions asking about customer loyalty tend to be concentrated in the 5-6 range. As a result, many respondents are classified as detractors (especially weak detractors). In addition, so-called loyal customers, who have a strong attachment to and high trust in the product, may intentionally give low scores to express their expectations of the product. Therefore, if respondents are classified solely based on the scores they give to questions asking about customer loyalty, respondents such as loyal customers who should actually be classified as promoters may end up being classified as neutrals or detractors. Another reason for this inability to properly classify respondents is that questions asking about customer loyalty are heavily dependent on the respondent's thoughts and feelings. For example, loyal customers who are reluctant to share their opinions and values ​​with others (sometimes referred to as introverted loyal customers) tend to give low scores on questions about customer loyalty. Therefore, it is highly likely that introverted loyal customers will not be classified as promoters. Indeed, introverted loyal customers are less likely to explicitly recommend the product they are evaluating than loyal customers who actively share their opinions and values ​​with others (sometimes referred to as extroverted loyal customers). On the other hand, regardless of whether they are introverted or extroverted loyal customers, the product they are evaluating tends to be deeply ingrained in their daily lives. Therefore, for example, if the product they are evaluating happens to appear in a photo or video they post on a social networking service (SNS) without intending to recommend it to others, introverted loyal customers are more likely to implicitly recommend the product they are evaluating to others. Therefore, it is desirable for loyal customers, regardless of whether they are introverted or extroverted loyal customers, to be classified as promoters.

[0054] One of the challenges with NPS is that it is difficult to intuitively understand the effects of NPS-based competitor comparisons or NPS-based improvement measures. One reason for this difficulty is that the score of the evaluation target alone does not allow one to identify the reason for the evaluation target's score. In particular, when the evaluation target is a conglomerate that operates across multiple industries, it is particularly difficult to intuitively understand the effects of NPS-based competitor comparisons or NPS-based improvement measures.

[0055] In contrast, the new evaluation index introduced in the system 10 according to the present embodiment is an evaluation index calculated based on the classification results of the evaluators 250 who evaluated the evaluation targets based on evaluation information including score information and comment information. For example, the information processing device 100 classifies the evaluation targets based on the scores of the evaluation targets indicated by the score information included in the evaluation information, and reclassifies the evaluation targets classified based on the scores of the evaluation targets based on comments related to the reasons for the scores of the evaluation targets indicated by the comment information included in the evaluation information. The new evaluation index introduced in the system 10 according to the present embodiment is an evaluation index calculated from the classification results in which the evaluators 250 are classified based not only on the scores of the evaluation targets but also on comments related to the reasons for the scores of the evaluation targets. Therefore, compared to conventional evaluation indexes calculated from the classification results in which the evaluators 250 are classified based only on the scores of the evaluation targets, the new evaluation index is calculated from the classification results in which the evaluation targets are more appropriately classified. Therefore, by classifying the evaluation targets based on evaluation information including score information and comment information, the system 10 according to the present embodiment can contribute to the introduction of evaluation indexes calculated from the classification results in which the evaluation targets are more appropriately classified than conventional evaluation indexes. The information processing device 100 calculates evaluation indices from classification results obtained by classifying evaluation targets based on evaluation information including score information and comment information. Therefore, the system 10 according to this embodiment can introduce evaluation indices calculated from classification results that more appropriately classify evaluation targets than conventional evaluation indices. Furthermore, the system 10 according to this embodiment identifies the reasons for the score of the evaluation target based on comments related to the reasons for the score of the evaluation target indicated by the comment information included in the evaluation information, and calculates an evaluation index for each identified reason for the score of the evaluation target. This allows the system 10 according to this embodiment to introduce evaluation indices that make it easier to intuitively understand competitive comparisons based on evaluation indices or the improvement effects of improvement measures based on evaluation indices, compared to conventional evaluation indices.

[0056] In particular, when the new evaluation index introduced in the system 10 according to this embodiment is the C-NPS, compared to the NPS, the C-NPS is calculated from the results of a classification that appropriately classifies evaluation targets, and is an evaluation index that makes it easy to intuitively understand competitive comparisons based on the evaluation index or the improvement effects of improvement measures based on the evaluation index. Therefore, the C-NPS is an evaluation index that solves the issues of the NPS. By implementing the C-NPS, the system 10 according to this embodiment can quantify customer loyalty with high reliability and introduce an evaluation index that makes it easy to intuitively understand competitive comparisons and the improvement effects of improvement measures. Therefore, by implementing the PDCA cycle based on the C-NPS, which includes planning (Plan) for selecting initiatives, execution (Do), measurement and evaluation (Check) for factor analysis, and measures and improvements (Action) for verifying the improvement effects, the system 10 according to this embodiment can contribute to the efficient and continuous improvement of companies. As a result, the system 10 according to this embodiment can contribute to the efficient and continuous development of companies.

[0057] 3 is an explanatory diagram for explaining an outline of a method for calculating evaluation indices. Here, an outline of a method for calculating C-NPS will be mainly explained. It is assumed that the evaluation target is a company 300.

[0058] The information processing device 100 classifies the evaluators 250 based on the scores of their answers to questions about customer loyalty toward the company 300, which are indicated by score information included in the evaluation information. In the example shown in Fig. 3, the information processing device 100 classifies A% of the evaluators 250 as "promoters," B% of the evaluators 250 as "neutrals," C% of the evaluators 250 as "weak detractors," and D% of the evaluators 250 as "strong detractors."

[0059] The information processing device 100 calculates the NPS from the classification result obtained by classifying the evaluators 250 based on the scores of their answers to a question about customer loyalty toward the company 300. In the example shown in FIG. 3, the information processing device 100 calculates NPS=A-(C+D).

[0060] The information processing device 100 classifies comments related to the reasons for the scores of the company 300, which are indicated by comment information included in the evaluation information, for example. The information processing device 100 classifies the comments by, for example, acquiring a prompt for classifying the comments, inputting the acquired prompt to the generation AI, and acquiring from the generation AI a classification result in which the comments are classified based on the prompt.

[0061] For example, the information processing device 100 classifies comments related to the reason for the score of the evaluation object into any of positive comments regarding the evaluation object, neutral comments regarding the evaluation object, negative comments regarding the evaluation object, and indifferent comments regarding the evaluation object. Note that positive comments regarding the evaluation object may be referred to as positive comments, neutral comments regarding the evaluation object may be referred to as neutral comments, negative comments regarding the evaluation object may be referred to as negative comments, and indifferent comments regarding the evaluation object may be referred to as indifferent comments.

[0062] A positive comment may be a comment that includes positive content regarding the evaluation target. A neutral comment may be a comment that includes neutral content regarding the evaluation target. A negative comment may be a comment that includes negative content regarding the evaluation target. An indifferent comment may be a comment that includes indifferent content regarding the evaluation target, or may be a comment that does not include any of positive content, neutral content, or negative content regarding the evaluation target.

[0063] When a comment related to the reason for the score of the evaluation target includes both positive and negative content regarding the evaluation target, the information processing device 100 may classify the comment as a neutral comment. The information processing device 100 may also classify the comment as either a positive comment or a negative comment.

[0064] For example, the information processing device 100 reclassifies the evaluator 250 who has been classified as a promoter, a neutral, or a detractor into a promoter, a neutral, a detractor, or an indifferent person based on the score of the answer to a question asking about customer loyalty toward the company 300 from the classification result of classifying comments related to the reasons for the score of the company 300. In the example shown in Fig. 3, the information processing device 100 reclassifies the evaluator 250 as follows.

[0065] The information processing device 100 reclassifies as a "promoter" an evaluator 250 who answers a question about customer loyalty toward the company 300 with a score of "9 to 10" and whose comments related to the reason for the score of the company 300 are "positive," including the content that "they gave me friendly advice about the pricing plan."

[0066] The information processing device 100 reclassifies, as a "promoter," an evaluator 250 who answers a question asking about customer loyalty to the company 300 with "7 to 8 points" and whose comment related to the reason for the score of the company 300 is "positive," including the content that "the rate plan is a good deal." The information processing device 100 reclassifies, as a "neutral," an evaluator 250 who answers a question asking about customer loyalty to the company 300 with "7 to 8 points" and whose comment related to the reason for the score of the company 300 is "neutral," including the content that "there are few line problems, but the rates are higher than other companies."

[0067] The information processing device 100 reclassifies as an "indifferent" evaluator 250 who answers a question asking about customer loyalty to the company 300 with "5 to 6 points" and whose comment related to the reason for the company 300's score is "indifferent" and includes the content "I don't really understand it myself, so I can't recommend it to others" as an "indifferent" evaluator. The information processing device 100 reclassifies as a "neutral" evaluator 250 who answers a question asking about customer loyalty to the company 300 with "5 to 6 points" and whose comment related to the reason for the company 300's score is "positive" and includes the content "The line is easy to connect to, so I feel quite safe" as an "indifferent" evaluator. The information processing device 100 reclassifies as a "neutral" evaluator 250 who answers a question asking about customer loyalty to the company 300 with "5 to 6 points" and whose comment related to the reason for the company 300's score is "neutral" and includes the content "There are no particular problems" as an "neutral" evaluator. The information processing device 100 reclassifies as a "critic" an evaluator 250 who answers a question about customer loyalty to the company 300 with "5 to 6 points" and whose comments related to the reason for the score of the company 300 are "negative" comments, including the content that "there is no benefit for long-term users."

[0068] The information processing device 100 reclassifies the evaluators 250 who answered "0 to 4 points" to the question asking about customer loyalty to the company 300 and whose comments related to the reason for the score of the company 300 were "neutral" comments, including the content that "communication quality is average," into "critics." The information processing device 100 reclassifies the evaluators 250 who answered "0 to 4 points" to the question asking about customer loyalty to the company 300 and whose comments related to the reason for the score of the company 300 were "negative" comments, including the content that "store staff are unfriendly," into "critics."

[0069] 3, the classification results obtained by the information processing device 100 reclassifying the evaluators 250 based on the classification results of the comments are as follows: The information processing device 100 reclassifies E'% of the evaluators 250 as "indifferent", A'% of the evaluators 250 as "promoters", B'% of the evaluators 250 as "neutral", and C'% of the evaluators 250 as "critics".

[0070] The information processing device 100 calculates the C-NPS from the classification result obtained by reclassifying the evaluators 250 based on the classification result of the comments, for example. The C-NPS may be an evaluation index that satisfies the relationship "C-NPS = (promoter rate after reclassification) - (detractor rate after reclassification)". In the example shown in FIG. 3, the information processing device 100 calculates C-NPS = A' - C'.

[0071] Here, we will consider the differences between the classification results obtained by classifying the evaluators 250 based on the scores of their responses to questions asking about their customer loyalty to the company 300 in the example shown in FIG. 3 and the classification results obtained by reclassifying the evaluators 250 based on comments related to the reasons for their scores to the company 300. Here, we will mainly consider three perspectives: the rate of promoters, the rate of detractors, and the inclusion of indifferent users. Note that classifying the evaluators 250 based on the scores of their responses to questions asking about their customer loyalty to the evaluation target may be referred to as NPS classification, and the classification results of NPS classification may be referred to as NPS classification results. Furthermore, reclassifying the evaluators 250 based on comments related to the reasons for their scores to the evaluation target may be referred to as C-NPS classification, and the classification results of C-NPS classification may be referred to as C-NPS classification results.

[0072] The promoter percentage in the NPS classification results is A%. On the other hand, the promoter percentage in the C-NPS classification results is A'%. Therefore, the promoter percentage in the C-NPS classification results is different from the promoter percentage in the NPS classification results. This means that the C-NPS classification can appropriately reclassify evaluators 250 who could not be classified as promoters in the NPS classification, such as loyal customers who intentionally gave low scores to express their expectations of the company 300, or introverted loyal customers, into promoters.

[0073] The percentage of detractors in the NPS classification results is C+D%. On the other hand, the percentage of detractors in the C-NPS classification results is C'%. Therefore, the percentage of detractors in the C-NPS classification results is different from the percentage of detractors in the NPS classification results. This means that the C-NPS classification can appropriately reclassify the evaluators 250 who answered 5-6 points to the question about customer loyalty toward Company 300, who would have been uniformly classified as detractors in the NPS classification, based on their comments related to the reasons for their score for Company 300.

[0074] The inventors discovered that indifferent users are less likely to make positive or negative comments about the target, while also less likely to switch from the target to its competitors. The tendencies of indifferent users do not match those of promoters, passive users, or detractors. Therefore, the C-NPS classification can reclassify the rater 250 into an indifferent user, who does not exist in the NPS classification. By introducing a new category of indifferent users, which did not exist in the NPS classification, the C-NPS classification results can not only calculate a C-NPS that can quantify customer loyalty with high reliability, but also generate data useful for determining the circulation policy of the PDCA cycle. As a result, the system 10 according to the embodiment shown in FIG. 3 can further contribute to the efficient and continuous development of companies.

[0075] Figure 4 is an explanatory diagram for explaining an example of comparing evaluation targets based on evaluation indexes. Here, an example of a competitive comparison of evaluation targets based on C-NPS will be mainly explained. The comparison targets for the competitive comparison are assumed to be telecommunications carriers A, B, and C.

[0076] For example, the information processing device 100 performs NPS classification and C-NPS classification for each of companies A, B, and C, thereby calculating the C-NPS for each of companies A, B, and C. To calculate the C-NPS for each of companies A, B, and C, the information processing device 100 identifies the reason for the score based on comments related to the reason for the score in the answer of the evaluator 250 to a question asking about customer loyalty toward the telecommunications carrier.

[0077] The reasons for the score include, for example, "fees," "customer contact points," "network," "economic sphere," "brand," and "other." "Fees" refers to matters related to the prices of the telecommunications carrier's products. "Customer contact points" refers to matters related to the contact points or interactions between customers and the telecommunications carrier. Examples of contact points or interactions between customers and the telecommunications carrier include customer service by the telecommunications carrier's store staff, and applications used by customers when checking contract details. "Network" refers to matters related to the telecommunications carrier's communication network. "Economic sphere" refers to matters related to electronic payment services and the like that accompany the telecommunications carrier's communication services. "Brand" refers to matters related to the telecommunications carrier's brand image. "Other" refers to matters that do not fall under any of "fees," "customer contact points," "network," "economic sphere," or "brand."

[0078] For example, the information processing device 100 calculates the C-NPS for each reason for the score by performing C-NPS classification for each of the identified reasons for the score for each of companies A, B, and C, and then calculates the C-NPS for each of companies A, B, and C by summing the C-NPS for each reason for the score. In the example shown in Figure 4, the information processing device 100 calculates the C-NPS for each of companies A, B, and C as follows.

[0079] The information processing device 100 calculates the C-NPS for Company A's "fees" as "-a1", the C-NPS for Company A's "customer contacts" as "+a2", the C-NPS for Company A's "network" as "+a3", the C-NPS for Company A's "economic sphere" as "+a4", the C-NPS for Company A's "brand" as "+a5", and the C-NPS for Company A's "other" as "+a6". The information processing device 100 then calculates the C-NPS for "Company A" as "-a1+a2+a3+a4+a5+a6=AA".

[0080] The information processing device 100 calculates the C-NPS for Company B's "fees" as "+b1", the C-NPS for Company B's "customer contacts" as "+b2", the C-NPS for Company B's "network" as "-b3", the C-NPS for Company B's "economic sphere" as "+b4", the C-NPS for Company B's "brand" as "+b5", and the C-NPS for Company B's "other" as "0". The information processing device 100 then calculates the C-NPS for "Company B" as "+b1+b2-b3+b4+b5+0=BB".

[0081] The information processing device 100 calculates the C-NPS for Company C's "fees" as "+c1", the C-NPS for Company C's "customer contacts" as "-c2", the C-NPS for Company C's "network" as "-c3", the C-NPS for Company C's "economic sphere" as "0", the C-NPS for Company C's "brand" as "+c5", and the C-NPS for Company C's "other" as "+c6". The information processing device 100 then calculates the C-NPS for "Company C" as "+c1-c2-c3+0+c5+c6=CC".

[0082] Here, we will explain an example of a competitive comparison using the C-NPS of companies A, B, and C shown in Figure 4, with company A as the comparison standard. Among the C-NPS for each reason for the score, the C-NPS for company A's "price," calculated as "-a1," is particularly lower than the C-NPS for company B's "price," calculated as "+b1," and the C-NPS for company C's "price," calculated as "+c1." On the other hand, among the C-NPS for each reason for the score, the C-NPS for company A's "network," calculated as "+a3," is particularly higher than the C-NPS for company B's "network," calculated as "-b3," and the C-NPS for company C's "network," calculated as "-c3." Therefore, as a result of the competitive comparison, it can be intuitively understood that company A's "price" is its weak point and its "network" is its strong point when competing in the market with companies B and C. Therefore, Company A can gain an advantage over Companies B and C in market competition by implementing a new business strategy that leverages its strengths while overcoming its weaknesses, such as by reviewing its pricing plans while maintaining the communication quality of its communication network. Furthermore, after implementing its new business strategy, Company A can intuitively grasp the improvement effects of the new business strategy by re-performing a competitive comparison using the C-NPS for each reason for the scores of Companies A, B, and C. Therefore, by calculating the C-NPS of the target to be evaluated by classifying the C-NPS for each reason for the scores of the evaluators 250 in response to questions asking about customer loyalty toward the target, the system 10 according to the embodiment shown in FIG. 4 can introduce the C-NPS, which, compared to the NPS, makes it easier to intuitively grasp the improvement effects of improvement measures based on evaluation indicators.

[0083] An example of a questionnaire is shown in outline in Figure 5. In Figure 5, an example of a questionnaire for calculating the C-NPS of a telecommunications carrier is shown.

[0084] The questionnaire includes, for example, questions for answering scores for evaluation objects. In the example of the questionnaire shown in Figure 5, the question (Q1) for answering scores for evaluation objects is, "Imagine that a family member or friend has asked you for advice about a mobile phone company. How likely would you be to recommend Company A? Please answer on a scale of 0 to 10. 10 points means you would most like to recommend them, and 0 points means you would least like to recommend them."

[0085] The questionnaire includes, for example, questions for responding in the form of comments about the reason for the score of the evaluation object. In the example of the questionnaire shown in Figure 5, the question (Q2) is, "Please tell us the details of why you scored Q1 as high as you did. If there is a specific episode, please also tell us the specific episode." This is a question for responding in the form of comments about the reason for the score of the evaluation object.

[0086] The questionnaire includes, for example, questions for responding with multiple choices about the reason for the score of the evaluation target. In the example of the questionnaire shown in FIG. 5, the question (Q3) reads, "From the options below, please select the option that corresponds to the reason for the score you gave to the evaluation target (multiple choices allowed)." This question (Q3) allows the evaluator 250 to select the option that corresponds to the reason for the score of the evaluation target from the options: "Price / Advantages and disadvantages for long-term users / Points and payment services / Discount campaigns / Campaigns for customers (distribution of goods, etc.) / Gift campaigns / Waiting time at the store / Explanation and suggestions from store staff / General customer service / Troubleshooting and repair response / Store experience / Call center / Device types and product lineup / Device prices / Device initial settings and usage / Network (radio waves, communication speed, etc.) / Wi-Fi (registered trademark) spots / Content services / Official advertising website / Comparisons and differences with other companies / Company image / Other / Other / Just because."

[0087] An example of a prompt is shown generally in Figure 6. Figure 6 shows an example of a prompt for categorizing comments by assigning labels to the comments related to the reasons for the scores of evaluators 250 in their responses to questions asking about customer loyalty to the carrier.

[0088] The prompt includes, for example, an "instruction" for instructing the generation AI. In the example of the prompt shown in Figure 6, "You are a comment analysis professional. Please assign labels to the comments in the input sentence according to the various conditions below." is an instruction for instructing the generation AI.

[0089] The prompt includes, for example, “labeling conditions” by the generation AI for comments related to the reasons for the score of the evaluator 250’s answer. In the example of prompts shown in Figure 6, "Pr1: Negative comments about the company's "fees (service fees and device fees)" and "fee services (plans, courses, optional services)" (high / expensive / not cheap / small discounts / limited plan options / no suitable plans / complicated plans / no family discounts / no discounts for bundled services with internet, etc.)", "Pr3: Positive comments about the company's "fees (service fees and device fees)" and "fee services (plans, courses, optional services)" (low / affordable / big discounts / large number of plan options / suitable plans / easy to understand plans / family discounts available / discounts for bundled services with internet / good value for switching / device fees are 1 yen, etc.)", "Pr2: Neutral or both positive and negative comments about the company's "fees (service fees and device fees)" and "fee services (plans, courses, optional services)", "Cs1: Negative comments about the company's "customer service (shops / stores or call centers / telephone service)" (poor service / takes a long time / can't get through on the phone, etc.)", "·Cs3: Positive comments about the company's "customer service (shop / store or call center / telephone response)"", "·Cs2: Neutral or both positive and negative comments about the company's "customer service (shop / store or call center / telephone response)", "·Nw1: Negative comments about the company's "communication environment (communication quality and coverage area)" (communication not connecting / cutting out / unusable / narrow / out of range, etc.)", "·Nw3: Positive comments about the company's "communication environment (communication quality and coverage area)", "·Nw2: "·Ec1: Negative comments about the company's "economic sphere (point redemption, QR code (registered trademark) payment, online shopping, credit card benefits, long-term user benefits)""; "·Ec3: Positive comments about the company's "economic sphere (point redemption, QR code payment, online shopping, credit card benefits, long-term user benefits)""; "·Ec2: Negative comments about the company's "economic sphere (point redemption,Neutral or both positive and negative comments about "QR code payments, online shopping, credit card benefits, long-term user benefits"; "Br1: Negative comments about the company's "brand image" (bad company image / not trusted / dishonest / lordly business, etc.); "Br3: Positive comments about the company's "brand image" (trusted / because it's a major company / because it's a major company / feelings of affection / good image / safe); "Br2: Neutral or both positive and negative comments about the company's "brand image"; "Ot1: Negative comments about the company's fees, fee services, customer service, communication environment, economic sphere, and other aspects other than the brand image (negative comments without specific details, etc.)); "Ot3: Positive comments about the company's fees, fee services, customer service, communication environment, economic sphere, and other aspects other than the brand image (positive comments without specific details, such as good / good, OK, etc.)" The conditions for labeling by the AI ​​are as follows: "·Ot2: Neutral or both positive and negative comments about content other than the company's fees, fee services, customer service, communication environment, economic zone, and brand image," "·Ot4: Comments similar to 'none in particular' (no anecdotes / no particular reason, etc.)," ​​"·Ot5: Comments similar to 'I don't know' (I don't really know / can't decide / just because / can't think of anything / just started using it, etc.)," ​​"·Ot6: Comments similar to 'I've been using it for a long time' (I've only ever signed up with their company / I don't know anything other than their company / I've never used other companies, etc.)," ​​"·Ot7: Comments similar to 'I don't recommend / can't recommend' (I don't recommend to others / it's up to each individual / it's personal freedom / I'm not knowledgeable enough to recommend / people around me use other companies / no one to recommend them, etc.)," ​​"·Ot8: Comments similar to 'I don't give good marks' (I don't give perfect marks / I hope for the future / I'm strict, etc.)," ​​and "·999: Comments whose meaning is unclear or incomprehensible."

[0090] The prompt includes, for example, an "input sentence" to which a label is to be assigned by the generation AI. In the example of the prompt shown in Figure 6, a comment related to the reason for the score of the evaluator 250's response to a question asking about customer loyalty to a telecommunications carrier is the input sentence to which a label is to be assigned by the generation AI.

[0091] Fig. 7 shows an example of a label assigned to a comment. Fig. 7 shows an example of a label when the generation AI assigns a label to a comment related to the reason for the score of the evaluator 250's answer to the question about customer loyalty to the telecommunications carrier in accordance with the prompt shown in Fig. 6. It should be noted that the categories of comments to be labeled are assumed to be "positive" comments, "neutral" comments, "negative" comments, and "indifferent" comments.

[0092] The label "Pr1" is a label assigned to comments whose reason is "fee / plan" and whose category is "negative." Examples of comments that are assigned the label "Pr1" include negative content about the fee amount (service fee or model price) and fee services (plan, course, optional services). The label "Pr2" is a label assigned to comments whose reason is "fee / plan" and whose category is "neutral." Examples of comments that are assigned the label "Pr2" include mixed positive and negative or neutral content about the fee amount (service fee or model price) and fee services (plan, course, optional services). Note that mixed positive and negative may mean that the comment contains both positive and negative content. The label "Pr3" is a label assigned to comments whose reason is "fee / plan" and whose category is "positive." Examples of comments labeled "Pr3" include "positive comments about the amount of fees (service fees and model fees)" and "fee services (plans, courses, and optional services)."

[0093] The label "Cs1" is assigned to comments whose reason is "customer contact (store / CS / Web)" and whose category is "negative." Examples of comments that are assigned the label "Cs1" include content that is "negative about customer service (customer service provided by the store / branch or call center / telephone)." The label "Cs2" is assigned to comments whose reason is "customer contact (store / CS / Web)" and whose category is "neutral." Examples of comments that are assigned the label "Cs2" include content that is "a mix of positive and negative comments or neutral about customer service (customer service provided by the store / branch or call center / telephone)." The label "Cs3" is assigned to comments whose reason is "customer contact (store / CS / Web)" and whose category is "positive." Examples of comments that are assigned the label "Cs3" include content that is "positive about customer service (customer service provided by the store / branch or call center / telephone)."

[0094] The label "Nw1" is a label assigned to comments whose reason is "network" and whose category is "negative." Examples of comments that are assigned the label "Nw1" include content that is "negative about the communication environment (communication quality and coverage area)." The label "Nw2" is a label assigned to comments whose reason is "network" and whose category is "neutral." Examples of comments that are assigned the label "Nw2" include content that is "a mix of positive and negative comments or neutral about the communication environment (communication quality and coverage area)." The label "Nw3" is a label assigned to comments whose reason is "network" and whose category is "positive." Examples of comments that are assigned the label "Nw3" include content that is "positive about the communication environment (communication quality and coverage area)."

[0095] The label "Ec1" is assigned to comments whose reason is "economic sphere" and whose category is "negative." Examples of comments labeled "Ec1" include content that is "negative toward the economic sphere (point redemption, QR code payment, online shopping, credit card benefits, and long-term user benefits)." The label "Ec2" is assigned to comments whose reason is "economic sphere" and whose category is "neutral." Examples of comments labeled "Ec2" include content that is "a mix of positive and negative opinions or neutral toward the economic sphere (point redemption, QR code payment, online shopping, credit card benefits, and long-term user benefits)." The label "Ec3" is assigned to comments whose reason is "economic sphere" and whose category is "positive." Examples of comments labeled "Ec3" include those that are "positive about the economic sphere (point redemption, QR code payments and internet shopping, credit card benefits, and benefits for long-term use)."

[0096] The label "Br1" is a label given to comments whose reason is "brand" and whose comment category is "negative." Examples of comments given the label "Br1" include content that is "negative toward the brand image." The label "Br2" is a label given to comments whose reason is "brand" and whose comment category is "neutral." Examples of comments given the label "Br2" include content that is "a mix of positive and negative comments or neutral toward the brand image." The label "Br3" is a label given to comments whose reason is "brand" and whose comment category is "positive." Examples of comments given the label "Br3" include content that is "positive toward the brand image."

[0097] The label "Ot1" is a label assigned to comments whose reason is "other" and whose category is "negative." Examples of comments that are assigned the label "Ot1" include content that is "negative toward anything other than the above (other than "price / plan," "customer contact point (store / CS / Web)," "network," "economic area," and "brand"). The label "Ot2" is a label assigned to comments whose reason is "other" and whose category is "neutral." Examples of comments that are assigned the label "Ot2" include content that is "a mix of positive and negative or neutral toward anything other than the above (other than "price / plan," "customer contact point (store / CS / Web)," "network," "economic area," and "brand"). The label "Ot3" is a label assigned to comments whose reason is "other" and whose category is "positive." Examples of comments labeled "Ot3" include those that are "positive about things other than the above (other than "price / plan," "customer touchpoint (store / CS / Web)," "network," "economic zone," and "brand").

[0098] The label "Ot4" is a label assigned to comments with the reason "none" and the comment category is "indifferent." Examples of comments assigned the label "Ot4" include "comments similar to 'none in particular' (no particular story / no particular reason, etc.)." The label "Ot5" is a label assigned to comments with the reason "none" and the comment category is "indifferent." Examples of comments assigned the label "Ot5" include "comments similar to 'I don't know' (I don't really understand / can't decide / just because / can't think of anything / just started using it, etc.)." The label "Ot6" is a label assigned to comments with the reason "none" and the comment category is "indifferent." Examples of comments assigned the label "Ot6" include "comments similar to 'I've been using it for a long time' (I've only ever had a contract with our company / I don't know anything other than our company / I've never used any other company, etc.)." The label "Ot7" is a label given to comments with "none" as the reason and in the "indifferent" category. Examples of comments that are given the label "Ot7" include comments similar to "do not recommend / not recommended" (do not recommend to others / it's up to each individual / personal freedom / not knowledgeable enough to recommend / people around me use other companies / no one to recommend, etc.). The label "Ot8" is a label given to comments with "none" as the reason and in the "indifferent" category. Examples of comments that are given the label "Ot8" include comments similar to "do not give good marks" (do not give perfect marks / hope for the future / being strict, etc.).

[0099] 8 shows an example of the churn rate of the evaluators 250. FIG. 8 shows an example of the churn rate of the evaluators 250 canceling their contracts with the company 300 within one year from when the evaluators 250 rated the company 300, for each combination of the category of the evaluators 250 and the category of comments related to the reasons for the scores of the answers of the evaluators 250 to questions asking about customer loyalty to the company 300.

[0100] The churn rate after one year for 250 raters whose category is "recommender" and whose comment category is "positive" is "CR r3 The rater category is "recommender" and the comment category is "neutral" comment. The cancellation rate after one year for 250 raters is "CR r2 The rate of cancellation after one year for 250 raters whose category is "recommender" and whose comment category is "negative" is "CR r1 The rate of cancellation after one year for 250 raters whose category is "recommender" and whose comment category is "indifferent" is "CR r4 %". Here, CR r1 %, CR r2 %, CR r3 % and CR r4 % is CR r4 %≒CR r3 % <CR r2 % <CR r1 The relationship of % is to be satisfied.

[0101] The cancellation rate after one year for 250 raters whose rater category is "Neutral" and whose comment category is "Positive" is "CR n3 The rater category is "Neutral" and the comment category is "Neutral" comment. The cancellation rate after one year for 250 raters is "CR n2 The cancellation rate after one year for 250 raters whose category is "Neutral" and whose comment category is "Negative" is "CR n1 The cancellation rate after one year for 250 raters whose category is "Neutral" and whose comment category is "Indifferent" is "CR n4 %". Here, CR n1 %, CR n2 %, CR n3 % and CR n4 % is CR n4 %≒CR n3 % <CR n2 % <CR n1 The relationship of % is to be satisfied.

[0102] The churn rate after one year for 250 raters whose rater category is "weak critics" and whose comment category is "positive" is "CR wc3 The rate of cancellation after one year for 250 raters whose category is "weak critics" and whose comment category is "neutral" is "CR wc2 The rate of cancellation after one year for 250 raters whose category of raters is "weak critics" and whose comment category is "negative" is "CR wc1 The rate of cancellation after one year for 250 raters whose category of raters is "weak critics" and whose comment category is "indifferent" is "CR wc4 %". Here, CR wc1 %, CR wc2 %, CR wc3 % and CR wc4 % is CR wc4 %≒CR wc3 % <CR wc2 % <CR wc1 The relationship of % is to be satisfied.

[0103] The churn rate after one year for 250 reviewers who are "strong critics" and whose comments are "positive" is "CR sc3 The rate of cancellation after one year for 250 reviewers whose category of reviewers is "strong critics" and whose comment category is "neutral" is "CR sc2 The rate of cancellation after one year for 250 reviewers whose category of reviewers is "strong critics" and whose comment category is "negative" is "CR sc1 The cancellation rate after one year for 250 reviewers whose category of reviewers is "strong critics" and whose comment category is "indifferent" is "CR sc4 %". Here, CR sc1 %, CR sc2 %, CR sc3 % and CR sc4 % is CR sc4 %≒CR sc3 % <CR sc2% <CR sc1 The relationship of % is to be satisfied.

[0104] Fig. 9 is an explanatory diagram for explaining an example of a method for calculating an evaluation index. Here, an example will be mainly explained in which the information processing device 100 performs C-NPS classification using an example of the churn rate of the evaluator 250 shown in Fig. 8. It is assumed that the evaluation target is a company 300.

[0105] The information processing device 100 classifies the evaluators 250 by, for example, performing NPS classification. In the example shown in Fig. 9, the information processing device 100 classifies A% of the evaluators 250 as "promoters," B% of the evaluators 250 as "passives," C% of the evaluators 250 as "weak detractors," and D% of the evaluators 250 as "strong detractors." In this case, the information processing device 100 calculates NPS = A - (C + D).

[0106] The information processing device 100 further classifies the evaluators 250 classified by the NPS classification, for example, based on comments related to the reasons for the scores of the company 300, which are indicated by comment information included in the evaluation information. In the example shown in Fig. 9, the information processing device 100 further classifies the evaluators 250 classified by the NPS classification as follows:

[0107] The information processing device 100 classifies A4% of the evaluators 250 as "recommenders" and "indifferent" comments, A3% of the evaluators 250 as "recommenders" and "positive" comments, and A2% of the evaluators 250 as "recommenders" and "neutral" comments. Note that the relationship A2% + A3% + A4% = A% is satisfied. The information processing device 100 classifies B4% of the evaluators 250 as "neutral" and "indifferent" comments, B3% of the evaluators 250 as "neutral" and "positive" comments, B2% of the evaluators 250 as "neutral" and "neutral" comments, and B1% of the evaluators 250 as "neutral" and "negative" comments. Note that the relationship B1% + B2% + B3% + B4% = B% is satisfied. The information processing device 100 classifies C4% of the raters 250 as "weak critics" and "indifferent" comments, C3% of the raters 250 as "weak critics" and "positive" comments, C2% of the raters 250 as "weak critics" and "neutral" comments, and C1% of the raters 250 as "weak critics" and "negative" comments. Note that the relationship C1% + C2% + C3% + C4% = C% is satisfied. The information processing device 100 classifies D4% of the raters 250 as "strong critics" and "indifferent" comments, D2% of the raters 250 as "strong critics" and "neutral" comments, and D1% of the raters 250 as "strong critics" and "negative" comments. Note that the relationship D1% + D2% + D4% = D% is satisfied.

[0108] The information processing device 100 reclassifies the evaluators 250 classified by the NPS classification into one of promoters, neutrals, detractors, and indifferents based on the churn rate of the evaluators 250 one year later for each combination of the evaluator 250 category and comment category. Here, the explanation will be continued assuming that the information processing device 100 reclassifies the evaluators 250 based on churn rate information, shown in FIG. 8 , indicating the churn rate at which the evaluator 250 cancels a contract with the company 300 within a predetermined period from when the evaluator 250 rated the company 300 being rated, for each combination of the evaluator 250 category and comment category. Examples of the predetermined period include one month, three months, six months, and one year. This churn rate information may also be referred to as first churn rate information.

[0109] For example, when the churn rate one year later of the evaluator 250 to be reclassified, which is a combination of the category of the evaluator 250 to be reclassified and the category of the comment indicated by the first churn rate information, is equal to or lower than the churn rate one year later of the evaluator 250 to be reclassified, which is a combination of the "recommender" and the "positive" comment indicated by the first churn rate information, the information processing device 100 reclassifies the evaluator 250 to be reclassified to "recommender". In an example shown in FIG. 9, the information processing device 100 n3 A3% of the combined "promoters" and "positive" comments were reclassified by 250 raters, and CR n3 %≦CR r3 250 raters who meet the B3% criteria for reclassification, combining "passive" and "positive" comments, will be reclassified as "promoters."

[0110] For example, when the churn rate one year later of the evaluator 250 to be reclassified, which is a combination of the category of the evaluator 250 to be reclassified and the category of the comment indicated by the first churn rate information, is higher than the churn rate one year later of the evaluator 250 with the combination of "recommender" and "positive" comment indicated by the first churn rate information, and is equal to or lower than the churn rate one year later of the evaluator 250 with the combination of "neutral" and "neutral" indicated by the first churn rate information, the information processing device 100 reclassifies the evaluator 250 to be reclassified to "neutral." In an example shown in FIG. 9, the information processing device 100 r3 % <CR r2 %≦CR n2 250% of the "promoters" and "neutral" comments that meet the A2% reclassification criteria. n2 250 reviewers who were reclassified as "Neutral" and "Neutral" comments, and 250 reviewers who were reclassified as "Neutral" and "Neutral" comments, r3 % <CR wc3 %≦CR n2 250 of the raters who meet the C3% reclassification target of the combination of "weak critics" and "positive" comments will be reclassified as "passives."

[0111] For example, if the churn rate one year later of the evaluator 250 to be reclassified, which is a combination of the category of the evaluator 250 to be reclassified and the category of the comment indicated by the first churn rate information, is higher than the churn rate one year later of the evaluator 250 to be reclassified, which is a combination of the "neutral" and "neutral" comments indicated by the first churn rate information, the information processing device 100 reclassifies the evaluator 250 to be reclassified to "detractor". In an example shown in FIG. 9, the information processing device 100 n2 % <CR n1 250 reviewers who met the B1% reclassification criteria for the combination of "neutral" and "negative" comments, CR n2 % <CR wc2 250% of the raters who met the C2% reclassification criteria for the combination of "weak critics" and "neutral" comments. n2 % <CR wc1250 raters to be reclassified for C1% of the combination of "weak critics" and "negative" comments that meet the CR n2 % <CR sc2 250 raters who met the D2% reclassification criteria for the combination of "strong critics" and "neutral" comments, and CR n2 % <CR wc1 250 raters who meet the D1% criteria for reclassification, combining "strong critics" and "negative" comments, will be reclassified as "critics."

[0112] The information processing device 100 may classify the evaluators 250 to be reclassified whose comment category is "indifferent" as "indifferent" regardless of the category of the evaluators 250 to be reclassified. In the example shown in Fig. 9, the information processing device 100 reclassifies A4% of the evaluators 250 to be reclassified who have a combination of "promoters" and "indifferent" comments, B4% of the evaluators 250 to be reclassified who have a combination of "neutrals" and "indifferent" comments, C4% of the evaluators 250 to be reclassified who have a combination of "weak critics" and "indifferent" comments, and D4% of the evaluators 250 to be reclassified who have a combination of "strong critics" and "indifferent" comments to "indifferent".

[0113] 9, the reclassification results of the information processing device 100 reclassifying the evaluators 250 based on the first churn rate information are as follows: The information processing device 100 reclassifies the evaluators 250 of A4+B4+C4+D4=E'% into "indifferents," reclassifies the evaluators 250 of A3+B3=A'% into "promoters," reclassifies the evaluators 250 of A2+B2+C3=B'% into "neutrals," and reclassifies the evaluators 250 of B1+C2+C1+D2+D1=C'% into "detractors."

[0114] The information processing device 100 calculates the C-NPS from the classification result obtained by reclassifying the evaluators 250 based on the first churn rate information, for example. In the example shown in Fig. 9, the information processing device 100 calculates, for example, C-NPS = A' - C'.

[0115] As shown in the example of the churn rate of the evaluator 250 shown in FIG. 8, the churn rate of indifferent people tends to be low regardless of the category of the evaluator 250. Therefore, the C-NPS may be an evaluation index that satisfies the relationship "C-NPS = (promoter rate after recategorization + indifferent rate after recategorization) - (detractor rate after recategorization)". Note that the indifferent rate is the rate of indifferent people who are indifferent to the evaluation target to the total number of evaluators who evaluated the evaluation target. Therefore, in the example shown in FIG. 9, the information processing device 100 may calculate C-NPS = (A' + E') - C'.

[0116] In order to gain an advantage over competitors in market competition, it is important for a company not only to develop new customers but also to maintain existing customers who bring a certain amount of revenue to the company. In particular, in industries with many competitors or industries that are shrinking in size overall, it is extremely difficult to develop new customers, so it is extremely important for a company to retain existing customers. Furthermore, the cost to a company of maintaining existing customers tends to be lower than the cost to a company of developing new customers. Therefore, it is desirable to develop a business strategy that takes into account the perspective of retaining existing customers.

[0117] In contrast, in the system 10 according to the embodiment shown in FIG. 9, the information processing device 100 reclassifies the NPS-classified evaluators 250 based on the first churn rate information. The churn rate is a parameter directly related to the retention of existing customers. Therefore, by reclassifying the NPS-classified evaluators 250 based not only on comments but also on churn rate, the system 10 according to the embodiment shown in FIG. 9 can contribute to the introduction of the C-NPS, which can generate useful data for developing business strategies that take into account the retention of existing customers. Furthermore, by calculating the C-NPS from the classification results obtained by the information processing device 100 reclassifying the NPS-classified evaluators 250 based on comments and churn rate, the system 10 according to the embodiment shown in FIG. 9 can introduce the C-NPS, which can generate useful data for developing business strategies that take into account the retention of existing customers. As a result, the system 10 according to the embodiment shown in FIG. 9 can further contribute to the realization of efficient and continuous business development.

[0118] FIG. 10 schematically illustrates another example of the churn rate of the evaluator 250. Here, it is assumed that the evaluator 250 is an indifferent person. FIG. 10 illustrates an example of the churn rate of the evaluator 250 canceling the contract with the company 300 within one year from when the evaluator 250 rated the company 300, for each label assigned to a comment related to the reason for the score of the evaluator 250's answer to a question asking about customer loyalty to the company 300 and for each category of options corresponding to the reason for the score of the evaluator 250's answer.

[0119] The churn rate one year after the indifferent users whose comments were labeled "Ot4" or "Ot5" as shown in Figure 7 and whose choice category was "reasonable" was "CR ot4,5w / The cancellation rate after one year for those who are indifferent and whose comments are labeled "Ot4" or "Ot5" and whose choice category is "No reason" is "CR ot4,5w / o %". The option category "with reason" means that the evaluator 250 selected an option that corresponds to the reason for the score of company 300 in response to a question that asks for an answer with multiple-choice options about the reason for the score of company 300, and the option category "no reason" means that the evaluator 250 did not select an option that corresponds to the reason for the score of company 300 in response to the question.

[0120] The label "Ot6" shown in Figure 7 is assigned to the comment, and the cancellation rate one year later for those who are uninterested and whose choice category is "reasonable" is "CR ot6w / The churn rate after one year for those who are indifferent and whose choice category is "No reason" is "CR ot6w / o %".

[0121] The label "Ot7" shown in Figure 7 is assigned to the comment, and the cancellation rate one year later for those who are uninterested and whose choice category is "reasonable" is "CR ot7w / The churn rate after one year for those who are indifferent and whose choice category is "No reason" is "CR ot7w / o %".

[0122] The label "Ot8" shown in Figure 7 is assigned to the comment, and the cancellation rate one year later for those who are uninterested and whose choice category is "reasonable" is "CR ot8w / The churn rate after one year for those who are indifferent and whose choice category is "No reason" is "CR ot8w / o %".

[0123] Fig. 11 is an explanatory diagram for explaining another example of a method for calculating an evaluation index. Here, an example will be mainly explained in which the information processing device 100 performs C-NPS classification using an example of the churn rate of the evaluator 250 shown in Fig. 11. It is assumed that the evaluation target is a company 300.

[0124] The information processing device 100 classifies the raters 250 by, for example, performing NPS classification and C-NPS classification. In the example shown in Fig. 11, the information processing device 100 reclassifies E'% of the raters 250 as "indifferents," A'% of the raters 250 as "promoters," B'% of the raters 250 as "passives," and C'% of the raters 250 as "detractors."

[0125] The information processing device 100 identifies the reason for the score of the company 300, for example, by assigning a label to a comment related to the reason for the score of the company 300, which is indicated by the comment information included in the evaluation information. In the example shown in FIG. 11 , the information processing device 100 identifies the reason for the score of the company 300 by assigning a label to a comment related to the reason for the score of the company 300. x The 250 comments of the raters who were reclassified as indifferent were given the label "Ot4" or "Ot5", and E' y The 250 comments from the raters who were reclassified as indifferent were given the label "Ot6", and E' z The 250 comments of the raters who were reclassified as indifferent are given the label "Ot8". x %+E' y %+E' z The relationship %=E'% is satisfied.

[0126] The information processing device 100 further classifies the evaluator 250 reclassified as an indifferent person based on, for example, the reason for the score of the company 300 identified by labeling the comment and the churn rate one year later of the evaluator 250 for each combination of options corresponding to the reason for the score of the company 300 indicated by the option information included in the evaluation information. Here, the explanation will be continued assuming that the information processing device 100 further classifies the evaluator 250 reclassified as an indifferent person based on churn rate information indicating the churn rate at which the evaluator 250 cancels a contract with the company 300 within a predetermined period from when the evaluator 250 evaluated the company 300, for each category of options and reasons for the score of the evaluation target identified from the comment, as shown in FIG. 10. Note that this churn rate information may be referred to as second churn rate information.

[0127] For example, when the churn rate one year later of the evaluator 250 who has been reclassified as an indifferent person for the combination of the reason for the score of the company 300 identified from the comments indicated by the second churn rate information and the option corresponding to the reason for the score of the company 300 is equal to or lower than the churn rate one year later of the evaluator 250 for the combination of "recommender" and "positive" comments indicated by the first churn rate information, the information processing device 100 classifies the evaluator 250 who has been reclassified as an indifferent person as an "indifferent person." In an example shown in FIG. 11, when the first churn rate information is the first churn rate information shown in FIG. 8, the churn rate one year later of the evaluator 250 for the combination of "recommender" and "positive" comments is equal to or lower than the churn rate one year later of the evaluator 250 for the combination of "recommender" and "positive" comments. r3 %”, the information processing device 100 ot4,5w / o %≦CR r3 % of the combination of label "Ot4" or "Ot5" and "No reason" w / o 250% of raters were reclassified as indifferent, and CR ot6w / %≦CR r3 % and CR ot6w / o %≦CR r3 % and satisfy the label "Ot6" E' y % of the raters who were reclassified as indifferent were reclassified as "indifferent."

[0128] For example, if the churn rate one year later of the evaluator 250 who has been reclassified as an indifferent person for the combination of the reason for the score of the company 300 identified from the comment indicated by the second churn rate information and the option corresponding to the reason for the score of the company 300 is higher than the churn rate one year later of the evaluator 250 for the combination of the "recommender" and "positive" comment indicated by the first churn rate information, the information processing device 100 classifies the evaluator 250 who has been reclassified as an indifferent person as a "neutral person." In an example shown in FIG. 11, the information processing device 100 r3 % <CR ot4,5w / % of the combination of label "Ot4" or "Ot5" and "reason" satisfy E' w / 250% of raters were reclassified as indifferent, and CR r3 % <CR ot8w / % and CR r3 % <CR ot8w / o % and satisfy the label "Ot8" E' z 250% of the raters who were reclassified as indifferent will be classified as "neutral."

[0129] In the example shown in FIG. 11, the information processing device 100 further classifies the evaluator 250 who has been reclassified as an indifferent person based on the second churn rate information as follows: w / o %+E' y Classify 250 of the raters (%=E''%) as "indifferent", reclassify 250 of the raters (A'%) as "promoters", and classify B'%+(E' w / %+E' z %) = Reclassify 250 of the B'% raters as "passives" and 250 of the C'% raters as "critics."

[0130] The information processing device 100 calculates the C-NPS from the classification result obtained by reclassifying the evaluators 250 based on the second churn rate information, for example. In the example shown in Fig. 11, the information processing device 100 calculates, for example, C-NPS = A' - C'.

[0131] Comment-style questions, which respondents can answer freely, can elicit more detailed responses from respondents than multiple-choice questions. On the other hand, comment-style questions place a greater burden on respondents than multiple-choice questions. Therefore, when comment-style questions are used to ask respondents about the reasons for their scores to company 300, some evaluators 250 find it troublesome to answer the comment-style questions and submit the questionnaire with no answer or only vague answers. As a result, these evaluators 250 tend to be reclassified as indifferent by the C-NPS classification. However, as mentioned above, the tendency of indifferent people does not match the tendency of promoters, neutrals, or detractors. Therefore, it is desirable to be able to determine whether evaluators 250 reclassified as indifferent are truly indifferent, or whether they simply found it troublesome to answer the questionnaire and are actually promoters, neutrals, or detractors rather than indifferent.

[0132] In contrast, in the system 10 according to the embodiment shown in FIG. 11, the information processing device 100 acquires evaluation information including option information indicating options corresponding to the reasons for the company 300's score, and further classifies the evaluator 250 who has been reclassified as an indifferent person based on the second churn rate information. By asking the evaluator 250 about the reasons for the company 300's score using multiple-choice questions, which place a smaller burden on the respondent than comment-style questions, and further classifying the evaluator 250 who has been reclassified as an indifferent person based on the option information and the second churn rate information, the system 10 according to the embodiment shown in FIG. 11 can more appropriately determine whether the evaluator 250 is truly indifferent. As a result, the system 10 according to the embodiment shown in FIG. 11 can introduce C-NPS, which can generate more useful data for developing business strategies, and can further contribute to the realization of efficient and continuous corporate growth.

[0133] 12 is a schematic diagram of another example of the system 10. Here, differences from the system 10 shown in FIG. 1 will be mainly described.

[0134] 12, the evaluation targets are a plurality of stores 350 of a company 300. The stores 350 may be branches or offices of the company 300.

[0135] For example, when the store 350 is visited by the evaluator 250, the information processing device 100 evaluates the store 350 visited by the evaluator 250 by conducting a questionnaire survey about the store 350 to the evaluator 250 after the evaluator 250 visits the store 350. For example, when a contract between the evaluator 250 and the company 300 is concluded at the store 350, the information processing device 100 evaluates the store 350 with which the contract between the evaluator 250 and the company 300 is concluded by conducting a questionnaire survey about the store 350 to the evaluator 250 on a regular or irregular basis.

[0136] 13 is an explanatory diagram for explaining another example of a method for calculating evaluation indexes. Here, differences from the above-mentioned method for calculating evaluation indexes will be mainly explained. It is assumed that the evaluation target is a store 350 of a company 300.

[0137] The information processing device 100 classifies the evaluators 250 by, for example, performing NPS classification. In the example shown in Fig. 13, the information processing device 100 classifies A% of the evaluators 250 as "promoters," B% of the evaluators 250 as "passives," and C% of the evaluators 250 as "detractors." In this case, the information processing device 100 calculates NPS=AC.

[0138] The information processing device 100 determines whether the reason for the score of the store 350 is attributable to at least one of the store 350 and the staff of the store 350, based on, for example, a comment in response to a comment-style question asking about the reason for the score of the store 350, which is indicated by comment information included in the evaluation information. For example, "The explanation was thorough and easy to understand," "The store was meticulously cleaned," "The telephone service was unpleasant," etc. are exemplified as examples of reasons for the score attributable to at least one of the store 350 and the staff of the store 350. On the other hand, for example, "There is no store nearby," "I am satisfied with the quality of the products," "The price is too high," "There is no one around me that I can recommend," no response, etc. are exemplified as examples of reasons for the score not attributable to either the store 350 or the staff of the store 350.

[0139] For example, the information processing device 100 extracts data of an evaluator 250 who has determined that the reason for the score is attributable to at least one of the store 350 and the crew of the store 350 as data to be used for calculating the C-NPS. On the other hand, the information processing device 100 does not extract data of an evaluator 250 who has determined that the reason for the score is not attributable to either the store 350 or the crew of the store 350 as data to be used for calculating the C-NPS. In this case, the information processing device 100 may remove the data of the evaluator 250.

[0140] In the example shown in FIG. 13, the information processing device 100 w / % of "promoters" data, B w / % of "neutral" data and C w / % of the data of “detractors” is extracted as data to be used for calculating the C-NPS. w / o % of "promoters" data, B w / o % of "neutral" data and C w / o Remove % of "detractors" data.

[0141] In the example shown in FIG. 13, after the information processing device 100 executes the data extraction process, A'% of the evaluators 250 are "promoters," B'% of the evaluators 250 are "neutrals," and C'% of the evaluators 250 are "detractors." In this case, "(promoters rate) - (detractors rate)" after the information processing device 100 executes the data extraction process is A' - C'. "(promoters rate) - (detractors rate)" after the data extraction process is executed may be an example of C-NPS.

[0142] For example, the information processing device 100 executes a data extraction process and then executes C-NPS classification to reclassify the evaluators 250 who were previously classified as recommenders, neutrals, or detractors into recommenders, neutrals, or detractors. For example, the information processing device 100 executes C-NPS classification for each reason for the score of the store 350 after removing data to be removed, to reclassify the evaluators 250 who were previously classified as recommenders, neutrals, or detractors into recommenders, neutrals, or detractors.

[0143] In the example shown in FIG. 13, the information processing device 100 reclassifies A''% of the evaluators 250 as "promoters," B''% of the evaluators 250 as "passives," and C''% of the evaluators 250 as "detractors." In this case, the information processing device 100 calculates C-NPS=A''-C''.

[0144] When the evaluation target is part of a company, such as a company's store, the reason for the evaluation target's score is not necessarily attributable to the evaluation target. Therefore, for example, when calculating evaluation indexes for each of a company's multiple stores to compare the multiple stores, if data from an evaluator whose score is not directly attributable to the store is used to calculate the evaluation index, it is impossible to achieve an appropriate comparison between the multiple stores using the evaluation indexes for each store. Therefore, when the evaluation target is part of a company, it is desirable to be able to prevent data from being used to calculate evaluation indexes for an evaluator whose score is not directly attributable to the evaluation target.

[0145] In contrast, in the system 10 according to the embodiment shown in FIG. 13, the information processing device 100 determines whether the reason for the score of the store 350 is attributable to at least one of the store 350 and the store's crew, based on comments related to the reason for the score indicated by the comment information included in the evaluation information. The information processing device 100 then extracts data on the evaluators 250 who determined that the reason for the score is attributable to at least one of the store 350 and the store's crew, and calculates the evaluation index using the extracted data. In this way, when the evaluation target is part of a company, the system 10 according to the embodiment shown in FIG. 13 can prevent the use of data on evaluators whose reason for the score is not directly attributable to the evaluation target in calculating the evaluation index. As a result, the system 10 according to the embodiment shown in FIG. 13 can realize appropriate comparisons between multiple stores using the evaluation indexes of each store.

[0146] 14 schematically illustrates an example of the functional configuration of the information processing device 100. The information processing device 100 includes an information storage unit 102, an acquisition unit 104, a classification unit 106, a calculation unit 108, an identification unit 110, a determination unit 112, a training data storage unit 114, a model generation unit 116, a model storage unit 118, and an output unit 120. Note that it is not essential that the information processing device 100 includes all of these components.

[0147] The information storage unit 102 stores various types of information. For example, the information storage unit 102 stores first churn rate information. For example, the information storage unit 102 stores second churn rate information.

[0148] The churn rates included in the first churn rate information and the second churn rate information may be determined based on the results of a questionnaire survey conducted in the past. The churn rates included in the first churn rate information and the second churn rate information may be updated each time churn rate data is accumulated by conducting a questionnaire survey.

[0149] The acquisition unit 104 acquires various types of information. For example, the acquisition unit 104 acquires the various types of information by receiving the various types of information via the network 20. The acquisition unit 104 may acquire the various types of information by an input unit included in the information processing device 100 accepting input of the various types of information. The acquisition unit 104 may store the acquired various types of information in the information storage unit 102.

[0150] The acquiring unit 104 acquires various information from, for example, the communication terminal 200. The acquiring unit 104 acquires various information from, for example, the collection device 400. The acquiring unit 104 may acquire various information from any other device.

[0151] The acquiring unit 104 acquires, for example, evaluation information, and a prompt for classifying comments related to the reasons for the score of the evaluation target, which are indicated by comment information included in the evaluation information.

[0152] The classification unit 106 classifies the evaluators 250 who evaluated the evaluation targets. The classification unit 106 classifies the evaluators 250 based on various information stored in the information storage unit 102, for example.

[0153] The classification unit 106 classifies the evaluators 250 based on, for example, the evaluation information. The classification unit 106 classifies the evaluators 250 based on, for example, score information included in the evaluation information. The classification unit 106 classifies the evaluators 250 by, for example, performing NPS classification.

[0154] For example, if the score of the evaluation target indicated by the score information is higher than a predetermined first score threshold, the classification unit 106 classifies the evaluator 250 as a recommender of the evaluation target. If the score of the evaluation target is lower than the first score threshold and higher than a predetermined second score threshold that is lower than the first score threshold, the classification unit 106 classifies the evaluator 250 as a neutral of the evaluation target. If the score of the evaluation target is lower than the second score threshold, the classification unit 106 classifies the evaluator 250 as a critic of the evaluation target.

[0155] The classification unit 106 may further classify the evaluator 250 classified as a critic of the evaluation target. For example, the classification unit 106 classifies the evaluator 250 as a weak critic of the evaluation target when the score of the evaluation target is lower than the second score threshold and higher than a predetermined third score threshold that is lower than the second score threshold. On the other hand, the classification unit 106 classifies the evaluator 250 as a strong critic of the evaluation target when the score of the evaluation target is lower than the third score threshold.

[0156] The classification unit 106 reclassifies the evaluators 250 classified based on the score information, for example, based on comment information included in the evaluation information. The classification unit 106 reclassifies the evaluators 250 classified by the scores of the evaluation targets indicated by the score information, for example, based on a classification result obtained by classifying comments indicated by the comment information according to predetermined classification criteria. The classification unit 106 reclassifies the evaluators 250 classified by the NPS classification, for example, by performing C-NPS classification.

[0157] The classification unit 106 classifies the comment indicated by the comment information into one of a positive comment, a neutral comment, a negative comment, and an indifferent comment, for example. The classification unit 106 classifies the comment by, for example, performing an analysis process on the comment. Note that the classification criterion for classifying the comment into one of a positive comment, a neutral comment, a negative comment, and an indifferent comment may be an example of the classification criterion for the comment. The comment may also be classified by a classification criterion different from the classification criterion for classifying the comment into one of a positive comment, a neutral comment, a negative comment, and an indifferent comment.

[0158] For example, if the comment contains positive content about the evaluation target, the classification unit 106 classifies the comment as a positive comment. For example, if the comment contains neutral content about the evaluation target, the classification unit 106 classifies the comment as a neutral comment. For example, if the comment contains negative content about the evaluation target, the classification unit 106 classifies the comment as a negative comment. For example, if the comment contains content that shows indifference about the evaluation target, or if the comment does not contain any of positive content, neutral content, and negative content about the evaluation target, the classification unit 106 classifies the comment as an indifferent comment.

[0159] For example, if the comment contains both positive and negative content regarding the evaluation target, the classification unit 106 classifies the comment as a neutral comment. If the comment contains both positive and negative content regarding the evaluation target, the classification unit 106 may classify the comment as either a positive comment or a negative comment.

[0160] For example, if the amount of positive content contained in the comment is greater than the amount of negative content contained in the comment, the classification unit 106 classifies the comment as a positive comment. On the other hand, if the amount of negative content contained in the comment is greater than the amount of positive content contained in the comment, the classification unit 106 classifies the comment as a negative comment.

[0161] For example, if the positive content of the comment contains stronger expressions than the negative content of the comment, the classification unit 106 classifies the comment as a positive comment. On the other hand, if the negative content of the comment contains stronger expressions than the positive content of the comment, the classification unit 106 classifies the comment as a negative comment.

[0162] The classification unit 106 classifies the comment into one of a positive comment, a neutral comment, a negative comment, and an indifferent comment, for example, by inputting the prompt acquired by the acquisition unit 104 into the generation AI and acquiring from the generation AI a classification result that classifies the comment based on the prompt. The generation AI classifies the comment by assigning to the comment one of a positive label indicating a positive comment, a neutral label indicating a neutral comment, a negative label indicating a negative comment, and an indifferent label indicating an indifferent comment.

[0163] "Pr3", "Cs3", "Nw3", "Ec3", "Br3", and "Ot3" shown in FIG. 7 are exemplified as examples of positive labels. "Pr2", "Cs2", "Nw2", "Ec2", "Br2", and "Ot2" shown in FIG. 7 are exemplified as examples of neutral labels. "Pr1", "Cs1", "Nw1", "Ec1", "Br1", and "Ot1" shown in FIG. 7 are exemplified as examples of negative labels. "Ot4", "Ot5", "Ot6", "Ot7", and "Ot8" shown in FIG. 7 are exemplified as examples of indifferent labels.

[0164] The classification unit 106 obtains the classification result of the comment, for example, by inputting a prompt to a generation AI installed in the information processing device 100. The classification unit 106 transmits the prompt to an external device via the network 20, and obtains the classification result of the comment by inputting the prompt to a generation AI installed in the external device.

[0165] The classification unit 106 reclassifies the evaluator 250 classified as a recommender of the evaluation target, a neutral person of the evaluation target, or a critic of the evaluation target based on the classification result of classifying the comment into, for example, a positive comment, a neutral comment, a negative comment, or an indifferent comment, into a recommender of the evaluation target, a neutral person of the evaluation target, a critic of the evaluation target, or an indifferent person who is indifferent to the evaluation target. For example, the classification unit 106 reclassifies the evaluator 250 for each combination of the category of the evaluator 250 classified based on the score information and the category of the comment classified based on the comment information.

[0166] For example, the classification unit 106 reclassifies the rater 250 of the combination of the recommender and positive comments of the evaluation target, the rater 250 of the combination of the recommender and neutral comments of the evaluation target, and the rater 250 of the combination of the neutral person and positive comments of the evaluation target into a recommender of the evaluation target. For example, the classification unit 106 reclassifies the rater 250 of the combination of the recommender and negative comments of the evaluation target, the rater 250 of the combination of the neutral person and neutral comments of the evaluation target, and the rater 250 of the combination of the weak critic and positive comments of the evaluation target into a neutral person of the evaluation target.

[0167] For example, the classification unit 106 reclassifies the rater 250 of the combination of a neutral person and negative comments of the rating target, the rater 250 of the combination of a weak critic and neutral comments of the rating target, the rater 250 of the combination of a weak critic and negative comments of the rating target, the rater 250 of the combination of a strong critic and positive comments of the rating target, the rater 250 of the combination of a strong critic and neutral comments of the rating target, and the rater 250 of the combination of a strong critic and negative comments of the rating target into a critic of the rating target. For example, the classification unit 106 reclassifies the rater 250 of the combination of a promoter and indifferent comments of the rating target, the rater 250 of the combination of a neutral person and indifferent comments of the rating target, the rater 250 of the combination of a weak critic and indifferent comments of the rating target, and the rater 250 of the combination of a strong critic and indifferent comments of the rating target into an indifferent person of the rating target.

[0168] The classification unit 106 reclassifies the evaluators 250 further based on the first churn rate information stored in the information storage unit 102. For example, the classification unit 106 reclassifies the evaluators 250 for each combination of the category of the evaluators 250 classified based on the score information and the category of the comments classified based on the comment information, further based on the first churn rate information.

[0169] For example, the classification unit 106 reclassifies the evaluator 250 as a promoter if the churn rate of the evaluator 250 within a predetermined period in the combination of the evaluator's 250 category and comment category indicated by the first churn rate information is equal to or lower than the churn rate of the evaluator 250 within the period in the combination of the recommender and positive comments indicated by the first churn rate information. For example, the classification unit 106 reclassifies the evaluator 250 as a neutral if the churn rate of the evaluator 250 within the period in the combination of the evaluator's 250 category and comment category is higher than the churn rate of the evaluator 250 within the period in the combination of the recommender and positive comments and is equal to or lower than the churn rate of the evaluator 250 within the period in the combination of the neutral and neutral categories indicated by the first churn rate information. For example, if the churn rate of the rater 250 in the period in the combination of the rater 250 category and the comment category is higher than the churn rate of the rater 250 in the period in the combination of the neutral person and the neutral comment, the classification unit 106 reclassifies the rater 250 as a detractor. For example, the classification unit 106 reclassifies the rater 250 whose comment category is indifferent comments as an indifferent person, regardless of the rater 250 category.

[0170] For example, the classification unit 106 reclassifies the evaluator 250 as a promoter if the churn rate of the evaluator 250 during the period in the combination of the evaluator's 250 category and the comment category is lower than a predetermined first churn rate threshold. For example, the classification unit 106 reclassifies the evaluator 250 as a passive if the churn rate of the evaluator 250 during the period in the combination of the evaluator's 250 category and the comment category is higher than the first churn rate threshold and lower than a predetermined second churn rate threshold that is higher than the first churn rate threshold. For example, the classification unit 106 reclassifies the evaluator 250 as a detractor if the churn rate of the evaluator 250 during the period in the combination of the evaluator's 250 category and the comment category is higher than the second churn rate threshold.

[0171] The calculation unit 108 calculates the evaluation index. For example, the calculation unit 108 calculates the evaluation index based on the classification result obtained by the classification unit 106 classifying the evaluators 250. For example, the calculation unit 108 calculates the evaluation index based on the classification result obtained by the classification unit 106 classifying the comments of the evaluators 250.

[0172] The calculation unit 108 calculates the evaluation index based on, for example, the NPS classification result obtained by the NPS classification performed by the classification unit 106. The calculation unit 108 calculates the evaluation index based on, for example, the C-NPS classification result obtained by the classification unit 106 performing C-NPS classification.

[0173] The calculation unit 108 calculates, for example, as an evaluation index, a difference between a recommender ratio, which is the ratio of the number of evaluators 250 classified by the classification unit 106 as recommenders of the evaluation target to the total number of evaluators 250 who evaluated the evaluation target, and a detractor ratio, which is the ratio of the number of evaluators 250 classified by the classification unit 106 as detractors of the evaluation target to the total number of evaluators 250 who evaluated the evaluation target. The calculation unit 108 calculates, for example, an NPS as an evaluation index.

[0174] The calculation unit 108 calculates, for example, as an evaluation index, a difference between a recommender ratio, which is the ratio of the number of evaluators 250 reclassified by the classification unit 106 into recommenders of the evaluation target to the total number of evaluators 250 who evaluated the evaluation target, and a detractor ratio, which is the ratio of the number of evaluators 250 reclassified by the classification unit 106 into detractors of the evaluation target to the total number of evaluators 250 who evaluated the evaluation target. The calculation unit 108 calculates, for example, C-NPS as an evaluation index.

[0175] The identification unit 110 identifies the reason for the score of the evaluation target indicated by the score information included in the evaluation information. The identification unit 110 identifies the reason for the score based on, for example, various information stored in the information storage unit 102.

[0176] The identification unit 110 identifies the reason for the score, for example, based on a comment related to the reason for the score indicated by comment information included in the evaluation information. The identification unit 110 identifies the reason for the score, for example, by performing an analysis process on the comment. The identification unit 110 identifies the reason for the score, for example, by assigning a label to the comment using a generation AI. Similar to the classification unit 106, the identification unit 110 may input the prompt acquired by the acquisition unit 104 to the generation AI and acquire the comment, to which a label has been assigned based on the prompt, from the generation AI, thereby identifying the reason for the score. The identification unit 110 may identify the reason for the score based on an option corresponding to the reason for the score indicated by option information included in the evaluation information.

[0177] The calculation unit 108, for example, calculates the difference between the promoter ratio and the detractor ratio for each reason for the score of the evaluation target identified by the identification unit 110. The calculation unit 108, for example, calculates the difference between the promoter ratio and the detractor ratio of the evaluation target by summing up the differences between the promoter ratio and the detractor ratio for each reason for the score of the evaluation target.

[0178] For example, when the classification unit 106 reclassifies the evaluator 250 as an indifferent person who is indifferent to the evaluation target, the classification unit 106 further classifies the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target, based on the option information included in the evaluation information. The option categories indicated by the option information include, for example, a category with a reason and a category without a reason. The category with a reason includes, for example, a category of options that are positive about the evaluation target, a category of options that are neutral about the evaluation target, a category of options that are negative about the evaluation target, and a category of options that are indifferent to the evaluation target.

[0179] For example, among the evaluators 250 who have been reclassified as indifferent people who are indifferent to the evaluation target, the classification unit 106 classifies the evaluators 250 who were classified as promoters of the evaluation target before being reclassified and who selected a positive option for the evaluation target, the evaluators 250 who were classified as promoters of the evaluation target before being reclassified and who selected a neutral option for the evaluation target, and the evaluators 250 who were classified as passive people for the evaluation target before being reclassified and who selected a positive option for the evaluation target as promoters of the evaluation target. For example, among the evaluators 250 who have been reclassified as indifferent people who are indifferent to the evaluation target, the classification unit 106 classifies the evaluators 250 who were classified as promoters of the evaluation target before being reclassified and who selected a negative option for the evaluation target, the evaluators 250 who were classified as passive people for the evaluation target before being reclassified and who selected a neutral option for the evaluation target, and the evaluators 250 who were classified as weak critics of the evaluation target before being reclassified and who selected a positive option for the evaluation target as passive people for the evaluation target.

[0180] For example, among the evaluators 250 who have been reclassified as indifferent people who are indifferent to the evaluation target, the classification unit 106 classifies the following as critics of the evaluation target: an evaluator 250 who was classified as a neutral person of the evaluation target before being reclassified and who selected a negative option regarding the evaluation target; an evaluator 250 who was classified as a weak critic of the evaluation target before being reclassified and who selected a neutral option regarding the evaluation target; an evaluator 250 who was classified as a weak critic of the evaluation target before being reclassified and who selected a negative option regarding the evaluation target; an evaluator 250 who was classified as a strong critic of the evaluation target before being reclassified and who selected a positive option regarding the evaluation target; an evaluator 250 who was classified as a strong critic of the evaluation target before being reclassified and who selected a neutral option regarding the evaluation target; and an evaluator 250 who was classified as a strong critic of the evaluation target before being reclassified and who selected a negative option regarding the evaluation target. For example, among the evaluators 250 who have been reclassified as indifferent people who are indifferent to the evaluation target, the classification unit 106 reclassifies the following evaluators 250, who were classified as promoters of the evaluation target before being reclassified and selected an option showing indifference to the evaluation target, the evaluators 250 who were classified as neutral people of the evaluation target before being reclassified and selected an option showing indifference to the evaluation target, the evaluators 250 who were classified as weak critics of the evaluation target before being reclassified and selected an option showing indifference to the evaluation target, and the evaluators 250 who were classified as strong critics of the evaluation target before being reclassified and selected an option showing indifference to the evaluation target, into the category of evaluators indifferent to the evaluation target.

[0181] The classification unit 106 further classifies the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target, for example, further based on the second churn rate information stored in the information storage unit 102. The classification unit 106 further classifies the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target, for each combination of the reason for the score of the evaluation target identified from the comment indicated by the comment information of the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target, and the category of options indicated by the option information of the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target, for example, further based on the second churn rate information.

[0182] For example, if the churn rate within a predetermined period of the evaluator 250 in the combination of the reason for the score and the category of the option identified from the comment, which is indicated by the second churn rate information, is equal to or lower than the churn rate within that period of the evaluator 250 in the combination of the recommender and the positive comment indicated by the first churn rate information, the classification unit 106 classifies the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target as an indifferent person. If the churn rate within that period of the evaluator 250 in the combination of the reason for the score and the category of the option identified from the comment is equal to or lower than the churn rate within that period of the evaluator 250 in the combination of the recommender and the positive comment, the classification unit 106 may classify the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target as a recommender.

[0183] For example, if the churn rate within the period of the evaluator 250 in the combination of the reason for the score and the category of the option identified from the comment is higher than the churn rate within the period of the evaluator 250 in the combination of the recommender and the positive comment and is equal to or lower than the churn rate within the period of the evaluator 250 in the combination of the neutral and neutral categories indicated by the first churn rate information, the classification unit 106 classifies the evaluator 250 reclassified as an indifferent person who is indifferent to the evaluation target as a neutral person. If the churn rate within the period of the evaluator 250 in the combination of the reason for the score and the category of the option identified from the comment is higher than the churn rate within the period of the evaluator 250 in the combination of the neutral and neutral comment, the classification unit 106 may classify the evaluator 250 reclassified as an indifferent person who is indifferent to the evaluation target as a detractor.

[0184] For example, if the churn rate within the period of the evaluator 250 for the combination of the reason for the score and the category of the option identified from the comment is lower than a predetermined third churn rate threshold, the classification unit 106 classifies the evaluator 250 reclassified as an indifferent person who is indifferent to the evaluation target as an indifferent person. If the churn rate within the period of the evaluator 250 for the combination of the reason for the score and the category of the option identified from the comment is lower than a third churn rate threshold, the classification unit 106 may classify the evaluator 250 reclassified as an indifferent person who is indifferent to the evaluation target as a recommender.

[0185] For example, if the churn rate within the period of the evaluator 250 for the combination of the reason for the score and the category of the option identified from the comment is higher than the third churn rate threshold and lower than a predetermined fourth churn rate threshold higher than the third churn rate threshold, the classification unit 106 classifies the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target as a neutral person. If the churn rate within the period of the evaluator 250 for the combination of the reason for the score and the category of the option identified from the comment is higher than the fourth churn rate threshold, the classification unit 106 may classify the evaluator 250 who has been reclassified as an indifferent person who is indifferent to the evaluation target as a critic.

[0186] The determination unit 112 determines whether the reason for the score of the evaluation object indicated by the score information included in the evaluation information is caused by the evaluation object. The determination unit 112 determines whether the reason for the score is caused by the evaluation object, for example, based on various information stored in the information storage unit 102.

[0187] The determination unit 112 determines whether the reason for the score is attributable to the evaluation target, for example, based on comment information included in the evaluation information. The determination unit 112 determines whether the reason for the score is attributable to the evaluation target, for example, based on option information included in the evaluation information.

[0188] For example, when the evaluation target is a store 350 of the company 300, the determination unit 112 determines whether the reason for the score is attributable to the evaluation target by determining whether the reason for the score is attributable to at least one of the store 350 and the crew of the store 350. For example, when the evaluation target is a product of the company 300, the determination unit 112 determines whether the reason for the score is attributable to the evaluation target by determining whether the reason for the score is attributable to the product.

[0189] For example, when the determination unit 112 determines that the reason for the score is due to the evaluation target, the classification unit 106 classifies the evaluator 250 based on the evaluation information including score information indicating the score. In this case, the classification unit 106 classifies the evaluator 250 based on the evaluation information, for example, by extracting the evaluation information. On the other hand, when the determination unit 112 determines that the reason for the score is not due to the evaluation target, the classification unit 106 does not classify the evaluator 250 based on the evaluation information. In this case, the classification unit 106 removes the evaluation information, for example.

[0190] The acquiring unit 104 may acquire learning data including, for example, comment data indicating comments related to the reasons for the score of the evaluation target and comment category data indicating the categories of the comments. The acquiring unit 104 may store the acquired learning data in the learning data storage unit 114.

[0191] The model generation unit 116 generates, from the comment data, a decision model that determines a comment category related to the reason for the score of the evaluation target indicated by the comment data. The model generation unit 116 generates the decision model by machine learning, for example, using a plurality of training data stored in the training data storage unit 114 as training data. The model generation unit 116 may store the generated decision model in the model storage unit 118.

[0192] The acquiring unit 104 may acquire a decision model. For example, the acquiring unit 104 may acquire a decision model similar to the decision model generated by the model generating unit 116. The acquiring unit 104 may store the acquired decision model in the model storage unit 118.

[0193] The classification unit 106 may classify the comment information included in the evaluation information using a determination model stored in the model storage unit 118. The classification unit 106 classifies the comment by, for example, using the determination model, determining from the comment information a comment category related to the reason for the score of the evaluation target indicated by the comment information.

[0194] The output unit 120 outputs various types of information. For example, the output unit 120 displays and outputs the various types of information on a display or the like provided in the information processing device 100. For example, the output unit 120 outputs the various types of information as audio to a speaker or the like provided in the information processing device 100. The output unit 120 may output the various types of information by transmitting the various types of information to an arbitrary communication device via the network 20.

[0195] The output unit 120 outputs, for example, the classification result obtained by the classification unit 106. The output unit 120 outputs, for example, the evaluation index calculated by the calculation unit .

[0196] The information processing device 100 may not include the training data storage unit 114, the model generation unit 116, and the model storage unit 118. In this case, the information processing device 100 may classify the comment information included in the evaluation information without using a determination model.

[0197] 15 is an explanatory diagram illustrating an example of the flow of processing in the system 10. Here, the description will be given assuming that the information processing device 100 is in a start state where it has not yet acquired evaluation information.

[0198] In step (sometimes abbreviated as S) 102, the acquisition unit 104 acquires evaluation information from the communication terminal 200. In S104, the classification unit 106 classifies the evaluators 250 into NPS categories based on the scores of the evaluation targets indicated by the score information included in the evaluation information acquired by the acquisition unit 104 in S102.

[0199] In S106, the classification unit 106 classifies comments related to the reasons for the scores of the evaluation target, which are indicated by the comment information included in the evaluation information acquired by the acquisition unit 104 in S102. In S108, the classification unit 106 C-NPS classifies the evaluators 250 classified in S104 based on the classification results of the comments in S106.

[0200] At S110, the identification unit 110 identifies the reason for the score of the evaluation target indicated by the score information based on the comment information. At S112, the calculation unit 108 calculates a C-NPS for each reason for the score of the evaluation target identified by the identification unit 110 at S110 based on the C-NPS classification results obtained by the classification unit 106 at S108. Thereafter, the calculation unit 108 calculates the C-NPS of the evaluation target by summing the C-NPS for each reason for the score of the evaluation target.

[0201] 16 schematically illustrates an example of the hardware configuration of a computer 1200 functioning as the information processing device 100. A program installed on the computer 1200 can cause the computer 1200 to function as one or more "units" of an apparatus according to the present embodiment, or can cause the computer 1200 to execute operations associated with the apparatus according to the present embodiment or one or more "units," and / or can cause the computer 1200 to execute a process according to the present embodiment or steps of the process. Such a program can be executed by the CPU 1212 to cause the computer 1200 to perform specific operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.

[0202] The computer 1200 according to this embodiment includes a CPU 1212, a RAM 1214, and a graphics controller 1216, which are interconnected by a host controller 1210. The computer 1200 also includes input / output units such as a communications interface 1222, a storage device 1224, a DVD drive 1226, and an IC card drive, which are connected to the host controller 1210 via an input / output controller 1220. The DVD drive 1226 may be a DVD-ROM drive, a DVD-RAM drive, or the like. The storage device 1224 may be a hard disk drive, a solid-state drive, or the like. The computer 1200 also includes a ROM 1230 and legacy input / output units such as a keyboard, which are connected to the input / output controller 1220 via an input / output chip 1240.

[0203] The CPU 1212 operates according to programs stored in the ROM 1230 and the RAM 1214, thereby controlling each unit. The graphics controller 1216 acquires image data generated by the CPU 1212 into a frame buffer or the like provided in the RAM 1214 or into the graphics controller itself, and causes the image data to be displayed on the display device 1218.

[0204] The communication interface 1222 communicates with other electronic devices via a network. The storage device 1224 stores programs and data used by the CPU 1212 in the computer 1200. The DVD drive 1226 reads programs or data from a DVD-ROM 1227 or the like and provides them to the storage device 1224. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.

[0205] The ROM 1230 stores therein a boot program or the like that is executed by the computer 1200 upon activation, and / or programs that depend on the hardware of the computer 1200. The input / output chip 1240 may also connect various input / output units to the input / output controller 1220 via a USB port, a parallel port, a serial port, a keyboard port, a mouse port, etc.

[0206] The programs are provided by a computer-readable storage medium such as a DVD-ROM 1227 or an IC card. The programs are read from the computer-readable storage medium, installed in the storage device 1224, RAM 1214, or ROM 1230, which are also examples of computer-readable storage media, and executed by the CPU 1212. Information processing described in these programs is read by the computer 1200, and causes cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be configured by implementing operations or processing of information in accordance with the use of the computer 1200.

[0207] For example, when communication is performed between the computer 1200 and an external device, the CPU 1212 may execute a communication program loaded into the RAM 1214 and instruct the communication interface 1222 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 1212, the communication interface 1222 reads transmission data stored in a transmission buffer area provided in the RAM 1214, the storage device 1224, the DVD-ROM 1227, or a recording medium such as an IC card, and transmits the read transmission data to the network, or writes reception data received from the network to a reception buffer area or the like provided on the recording medium.

[0208] Furthermore, the CPU 1212 may cause all or a necessary portion of a file or database stored in an external recording medium such as the storage device 1224, the DVD drive 1226 (DVD-ROM 1227), an IC card, etc. to be read into the RAM 1214, and may perform various types of processing on the data on the RAM 1214. The CPU 1212 may then write back the processed data to the external recording medium.

[0209] Various types of information, such as various types of programs, data, tables, and databases, may be stored on the recording medium and may undergo information processing. The CPU 1212 may perform various types of processing on data read from the RAM 1214, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequences of the programs, and write the results back to the RAM 1214. The CPU 1212 may also search for information in a file, database, etc. on the recording medium. For example, if multiple entries, each having an attribute value of a first attribute associated with an attribute value of a second attribute, are stored on the recording medium, the CPU 1212 may search for an entry whose attribute value of the first attribute matches a specified condition from among the multiple entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0210] The above-described programs or software modules may be stored in a computer-readable storage medium on or near the computer 1200. A recording medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can also be used as a computer-readable storage medium, thereby providing the programs to the computer 1200 via the network.

[0211] The blocks in the flowcharts and block diagrams in the present embodiments may represent stages of a process in which an operation is performed or "parts" of an apparatus responsible for performing the operation. Particular stages and "parts" may be implemented by dedicated circuitry, programmable circuitry provided with computer-readable instructions stored on a computer-readable storage medium, and / or a processor provided with computer-readable instructions stored on a computer-readable storage medium. The dedicated circuitry may include digital and / or analog hardware circuits, including integrated circuits (ICs) and / or discrete circuits. The programmable circuitry may include reconfigurable hardware circuits, such as field programmable gate arrays (FPGAs) and programmable logic arrays (PLAs), including AND, OR, XOR, NAND, NOR, and other logical operations, flip-flops, registers, and memory elements.

[0212] A computer-readable medium may include any tangible device capable of storing instructions that are executed by a suitable device, such that the computer-readable medium having instructions stored thereon comprises an article of manufacture containing instructions that can be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of computer-readable media may include electronic, magnetic, optical, electromagnetic, and semiconductor storage media. More specific examples of computer-readable media may include floppy disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray disc, memory stick, integrated circuit card, and the like.

[0213] The computer readable instructions may include either assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages ​​such as the “C” programming language or similar programming languages.

[0214] The computer-readable instructions may be provided to a processor or programmable circuit of a programmable data processing device, such as a computer, locally or via a wide area network (WAN) such as a local area network (LAN) or the Internet, and the computer-readable instructions may be executed to create means for performing the operations specified in the flowcharts or block diagrams. Here, the computer may be a personal computer (PC), a tablet computer, a smartphone, a workstation, a server computer, a general-purpose computer, a special-purpose computer, or the like, or may be a computer system in which multiple computers are connected. Such a computer system in which multiple computers are connected is also called a distributed computing system, and is a broad definition of computer. In a distributed computing system, the multiple computers collectively execute a program by each executing a portion of the program and passing data between the computers as needed during program execution.

[0215] Examples of processors include a computer processor, a central processing unit (CPU), a processing unit, a microprocessor, a digital signal processor, a controller, a microcontroller, etc. A computer may have one or more processors. In a multiprocessor system with multiple processors, each processor executes a portion of a program and passes data between processors as needed during program execution, allowing the multiple processors to collectively execute the program. For example, in multitasking, each of the multiple processors may execute a portion of each task in small chunks by switching tasks at time slice intervals. In this case, which portion of a program each processor executes changes dynamically. Which portion of a program each of the multiple processors executes may also be statically determined by multiprocessor-aware programming.

[0216] This invention can contribute to the introduction of evaluation indices calculated from classification results that more appropriately classify evaluation targets compared with conventional evaluation indices, and can therefore contribute to the achievement of Goal 9 of the Sustainable Development Goals (SDGs), "Build resilient infrastructure, promote inclusive and sustainable industrialization, and promote innovation and infrastructure."

[0217] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the claims that such modifications and improvements can also be included within the technical scope of the present invention.

[0218] It should be noted that the execution order of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and that the processes can be performed in any order unless the output of a previous process is used in a later process. Even if the operational flow in the claims, specifications, and drawings is described using "first," "next," etc. for convenience, this does not mean that the processes must be performed in this order. [Explanation of symbols]

[0219] 10 system, 20 network, 100 information processing device, 102 information storage unit, 104 acquisition unit, 106 classification unit, 108 calculation unit, 110 identification unit, 112 judgment unit, 114 learning data storage unit, 116 model generation unit, 118 model storage unit, 120 output unit, 200 communication terminal, 250 evaluator, 300 company, 350 store, 400 collection device, 1200 computer, 1210 host controller, 1212 CPU, 1214 RAM, 1216 graphics controller, 1218 display device, 1220 input / output controller, 1222 communication interface, 1224 storage device, 1226 DVD drive, 1227 DVD-ROM, 1230 ROM, 1240 input / output chip

Claims

1. an acquisition unit that acquires evaluation information including score information indicating a score of an evaluation object and comment information indicating a comment related to a reason for the score of the evaluation object; a classification unit that classifies the evaluator who evaluated the evaluation target as a recommender of the evaluation target when the score of the evaluation target indicated by the score information is higher than a predetermined first score threshold, classifies the evaluator as a neutral of the evaluation target when the score of the evaluation target is lower than the first score threshold and higher than a predetermined second score threshold that is lower than the first score threshold, and classifies the evaluator as a critic of the evaluation target when the score of the evaluation target is lower than the second score threshold; Equipped with The classification unit reclassifies the evaluator who has been classified into one of a recommender of the evaluation target, a neutral person of the evaluation target, and a critic of the evaluation target, into one of a recommender of the evaluation target, a neutral person of the evaluation target, a critic of the evaluation target, and an indifferent person who is indifferent to the evaluation target, based on a classification result obtained by classifying the comments indicated by the comment information according to a predetermined classification criterion. Information processing device.

2. 2. The information processing device according to claim 1, wherein the classification unit reclassifies the evaluators who have been classified as recommenders of the object to be evaluated, neutral people of the object to be evaluated, or critics of the object to be evaluated into recommenders of the object to be evaluated, neutral people of the object to be evaluated, critics of the object to be evaluated, or indifferent people who are indifferent to the object to be evaluated, based on a classification result in which the comments indicated by the comment information are classified into positive comments about the object to be evaluated, neutral comments about the object to be evaluated, negative comments about the object to be evaluated, or comments that are indifferent to the object to be evaluated.

3. A calculation unit that calculates the difference between a recommender ratio, which is the ratio of the number of evaluators who have been reclassified by the classification unit into recommenders of the evaluation target to the total number of evaluators who have evaluated the evaluation target, and a critic ratio, which is the ratio of the number of evaluators who have been reclassified by the classification unit into critics of the evaluation target to the total number of evaluators who have evaluated the evaluation target. The information processing device according to claim 1 , further comprising:

4. an identification unit that identifies the reason for the score of the evaluation target based on the comment information; Furthermore, the calculation unit calculates a difference between the promoter ratio and the detractor ratio for each reason for the score of the evaluation target. The information processing device according to claim 3 .

5. The acquisition unit further acquires a prompt for classifying the comment; The classification unit inputs the prompt to a generation AI and classifies the comment by obtaining the classification result, which classifies the comment based on the prompt, from the generation AI.

3. The information processing device according to claim 1 or 2.

6. An information storage unit that stores, for each combination of the evaluator category and the comment category, cancellation rate information indicating the cancellation rate at which the evaluator cancels a contract with the company within a predetermined period from when the evaluator evaluated the company being evaluated. Furthermore, The classification unit reclassifies the evaluators further based on the churn rate information stored in the information storage unit.

3. The information processing device according to claim 1 or 2.

7. the acquiring unit acquires the evaluation information further including option information indicating an option corresponding to a reason for the score of the evaluation target, When the classification unit reclassifies the evaluator into an indifferent person who is indifferent to the evaluation target, the classification unit further classifies the evaluator who has been reclassified into an indifferent person who is indifferent to the evaluation target based on the option information.

3. The information processing device according to claim 1 or 2.

8. an information storage unit that stores cancellation rate information indicating the cancellation rate at which the evaluator cancels a contract with the company within a predetermined period from when the evaluator evaluated the company to be evaluated, for each category of the options and the reason for the score of the evaluation target identified from the comment information; an identification unit that identifies the reason for the score of the evaluation target based on the comment information; Furthermore, The classification unit further classifies the evaluators who have been reclassified into indifferent people who are indifferent to the evaluation subject, based on the cancellation rate information stored in the information storage unit. The information processing device according to claim 7 .

9. a determination unit that determines, based on the comment information, whether the reason for the score of the store of the company to be evaluated, which is indicated by the score information, is attributable to at least one of the store and the store's crew. Furthermore, the classification unit classifies the evaluator based on the evaluation information when the determination unit determines that the reason for the score is attributable to at least one of the store and the crew.

3. The information processing device according to claim 1 or 2.

10. On the computer, an acquisition step of acquiring evaluation information including score information indicating a score of an evaluation object and comment information indicating a comment related to a reason for the score of the evaluation object; a classification procedure for classifying the evaluator who evaluated the evaluation object as a promoter of the evaluation object when the score of the evaluation object indicated by the score information is higher than a predetermined first score threshold, classifying the evaluator as a neutral of the evaluation object when the score of the evaluation object is lower than the first score threshold and higher than a predetermined second score threshold that is lower than the first score threshold, and classifying the evaluator as a critic of the evaluation object when the score of the evaluation object is lower than the second score threshold; a reclassification step of reclassifying the evaluator, who has been classified into one of a recommender of the evaluation target, a neutral person of the evaluation target, and a critic of the evaluation target, into one of a recommender of the evaluation target, a neutral person of the evaluation target, a critic of the evaluation target, and an indifferent person who is indifferent to the evaluation target, based on a classification result obtained by classifying the comments indicated by the comment information according to a predetermined classification criterion; A program to execute.

11. 1. A computer-implemented information processing method, comprising: an acquisition step of acquiring evaluation information including score information indicating a score of an evaluation object and comment information indicating a comment related to a reason for the score of the evaluation object; a classification step of classifying the evaluator who evaluated the evaluation object based on the evaluation information by classifying the evaluator who evaluated the evaluation object as a promoter of the evaluation object when the score of the evaluation object indicated by the score information is higher than a predetermined first score threshold, classifying the evaluator as a neutral of the evaluation object when the score of the evaluation object is lower than the first score threshold and higher than a predetermined second score threshold that is lower than the first score threshold, and classifying the evaluator as a critic of the evaluation object when the score of the evaluation object is lower than the second score threshold; a reclassification step of reclassifying the evaluator, who has been classified into one of a recommender of the evaluation target, a neutral person of the evaluation target, and a critic of the evaluation target, into one of a recommender of the evaluation target, a neutral person of the evaluation target, a critic of the evaluation target, and an indifferent person who is indifferent to the evaluation target, based on a classification result obtained by classifying the comments indicated by the comment information according to a predetermined classification criterion; An information processing method comprising:

Images