JP7297137B1 - Program, method and system - Google Patents

Abstract

A system capable of providing multifaceted information about potential customers is provided. According to a program of the present invention, a processor receives an input of an evaluation area to be evaluated; a management server that manages a search history of a user; aggregated by referring to past search histories and action histories, and including information indicating the percentage of each area actually visited during the evaluation period among the user groups who searched for the evaluation area during the evaluation period. , a step of acquiring from a management server, a step of determining a customer who has not visited an evaluation area in an evaluation period as a non-visiting customer among customers included in the acquired population data, and information related to the determined non-visiting customer are aggregated with reference to customer information managed in association with customer identification information. [Selection drawing] Fig. 10

Description

The present disclosure relates to programs, methods, and systems.

Conventionally, the customer's location information and search history are used to estimate the target of the customer's interest, and the target is used for various promotions.
Patent Literature 1 below discloses, as such a system, a system that evaluates the customer's degree of interest in an evaluation target item by associating the customer's stay time at a predetermined location with a search history. .

JP 2020-190964 A

However, in the conventional system, although it is possible to evaluate customers who have a high degree of interest in the evaluation target item, evaluation of potential customers who were temporarily interested in the evaluation target item but then lost interest in the evaluation target item. could not do enough.
Therefore, it has been desired to provide multifaceted information about potential customers by using the user's action history and search history.

SUMMARY OF THE INVENTION An object of the present invention is to provide a system capable of providing multifaceted information about potential customers.

A program of the present invention is provided with a computer having a processor and is used in a system for analyzing customer behavior, wherein the program comprises a step of receiving an input of an evaluation area to be evaluated, and a user search process. A step of designating an evaluation area for a management server that manages histories, and referring to past search histories and action histories in the management server, totaling, and among users who searched for the evaluation area during the evaluation period a step of acquiring from a management server population data including information indicating the percentage of each area actually visited during the evaluation period; a step of determining a non-visiting customer as a non-visiting customer; and a step of aggregating information related to the determined non-visiting customer by referring to customer information managed in association with customer identification information.

According to the present invention, it is possible to provide multifaceted information about potential customers.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows the outline|summary of the customer behavior analysis system which concerns on this embodiment. 2 is a block diagram showing a hardware configuration of an evaluator terminal shown in FIG. 1; FIG. It is a figure which shows the hardware constitutions of the action-analysis server shown in FIG. It is a figure which shows the functional structure of the behavior-analysis server shown in FIG. 2 is a diagram showing a hardware configuration of a customer information management server shown in FIG. 1; FIG. It is a figure which shows the example of a customer database. It is a figure which shows the specific example of an action log database. It is a figure which shows the specific example of a search history database. BRIEF DESCRIPTION OF THE DRAWINGS It is a figure explaining the first half of the outline|summary of embodiment of a customer behavior analysis system. It is a figure explaining the latter half of the outline|summary of embodiment of a customer behavior analysis system. It is a figure explaining the 1st example of the process by a customer behavior analysis system. It is a figure explaining the 2nd example of the process by a customer behavior analysis system. It is a figure explaining the determination example of a visiting customer and a non-visiting customer by a customer behavior analysis system. It is a figure explaining processing of keyword analysis by a customer behavior analysis system. It is a figure explaining the processing of the area analysis by a customer behavior analysis system. 16A is a diagram showing the ratio of visiting customers and non-visiting customers, FIG. 16B is a diagram showing the distribution of customer groups of visiting customers and non-visiting customers, and FIG. 16D is a diagram showing the ranking of search purpose keywords of visiting customers and non-visiting customers, and FIG. 16D is a diagram showing the ranking of visited destinations of visiting customers and non-visiting customers. It is a figure which shows the 2nd example of the output image of a customer behavior analysis system. It is a figure which shows the 3rd example of the output image of a customer behavior analysis system. It is a figure which shows the 4th example of the output image of a customer behavior analysis system.

An embodiment of the present invention will be described in detail below with reference to the drawings. In the drawings for describing the embodiments, in principle, the same constituent elements are denoted by the same reference numerals, and repeated description thereof will be omitted.

(1) Overview of Customer Behavior Analysis System 1 An overview of a customer behavior analysis system 1 (hereinafter simply referred to as system 1) according to the present embodiment will be described below.
The system 1 analyzes the search history of users using terminal devices such as mobile terminals and the behavior history of customers who have mobile terminals, thereby performing behavior analysis such as totaling customer behavior results for evaluation targets. is.

Specifically, for example, the system 1 is used in attracting tourism. The system 1 refers to a search history during a predetermined evaluation period, designates a user who has searched an evaluation target area as a customer candidate, and refers to the behavior history of this customer candidate using the location information of a mobile terminal, etc., to obtain a customer. Categorize the candidates into non-visiting customers and visiting customers.
・Non-visit customers: Among customer candidates, customers who have not visited the evaluation target area (potential customers)
・ Visiting customers: Among customer candidates, customers who visited the evaluation target area (acquired customers)
That is, the system 1 aggregates information on non-visiting customers and visiting customers and statistically analyzes the behavior of both to provide information suggesting effective measures for attracting customers to the evaluation target area.

Here, the concept of the evaluation period will be explained.
In the system 1, the evaluation period of the search history and action history is set to the past three months from the time of evaluation as the starting date.
Note that the evaluation period can be changed arbitrarily. At this time, the start date of the evaluation period may be different from the evaluation time.

Next, the classification of the evaluation area to be evaluated by the system 1 will be described.
Examples of classification of the evaluation area that can be evaluated by the system 1 include the following.
・Area specified by an administrative division such as the name of a prefecture or municipality ・Human resources (historic sites, shrines and temples, castles, gardens, parks, historical landscapes, regional landscapes, annual events, historical buildings, modern buildings) , museums, art galleries, etc.) ・Area specified by the name of natural resources (mountains, plateaus, wetlands, lakes, valleys, waterfalls, rivers, coasts, capes, islands, rocks, caves, etc.) Areas identified by the names of tourist areas that use natural resources (hot spring resorts, ski resorts, campgrounds, beaches) ・Downtowns (Susukino (Sapporo), Kabukicho (Tokyo), Nakasu (Fukuoka), Minami (Osaka) ), etc.) ・Area identified by the name of various facilities (amusement parks, theme parks, shopping centers, amusement centers, etc.) Note that the categories of evaluation areas are not limited to these examples. , a segment capable of identifying tourism resources having a certain ability to attract customers may be set as the evaluation area.
The configuration of such system 1 will be described below.

FIG. 1 is a block diagram showing the configuration of the system 1 of this embodiment.
As shown in FIG. 1, the system 1 includes a customer terminal 10, an evaluator terminal 20, a behavior analysis server 30, a customer information management server 40, a customer database (DB) 50, and a behavior history database (DB) 60. , and a search history database (DB) 70 .
The customer terminal 10, the evaluator terminal 20, the behavior analysis server 30, the customer information management server 40, the customer DB 50, the behavior history DB 60, and the search history DB 70 are connected to communicate with each other via a network (for example, the Internet or an intranet) 80. ing.

The customer terminal 10 is an information processing device such as a mobile phone carried by the customer. The customer terminal 10 communicates with a GPS satellite, which is the positioning device 81 , so that the GPS system identifies the position information of the customer terminal 10 . The location information of the specified customer terminal 10 is accumulated in the location information database. The GPS system continuously acquires position information and accumulates it in a position information database along a predetermined sampling period.

The positioning device is not limited to the GPS system, and various positioning systems that communicate with a communication device while moving and acquire the position information of the customer terminal 10, which is a communication device, can be adopted.
For example, other mobile positioning devices include GNSS (Global Navigation Satellite System) and RNSS (Regional Navigation Satellite System) other than GPS.

As the positioning device 81, a base station (fixed positioning device) that performs mobile communication with the customer terminal 10 may be employed. In this case, the customer terminal 10 is a communication device that is carried by the customer and can be called a mobile terminal. Examples of such mobile terminals include the following in addition to mobile phones.
・Wifi (registered trademark) beacons carried by customers ・Mobile antennas used for radar positioning ・In-vehicle car navigation terminals compatible with ITS Spot (DSRC) ・Communication terminals using Bluetooth (registered trademark) A mobile terminal may be any other device that communicates while moving.

Here, the fixed positioning device refers to a device that is fixed to a fixed position and performs communication, preferably wireless communication, with a communication device. positioning equipment for communication, positioning equipment for radar positioning, positioning equipment for ITSspot, etc.

The customer terminal 10 may include various computers such as a smart phone, a tablet terminal, a wearable device (eg, smart watch or smart glasses), etc., as devices that perform wireless communication with the positioning device 81 .
In this embodiment, as the customer terminal 10, a smart phone that performs wireless communication with a GPS satellite, which is the positioning device 81, will be described as an example.

The evaluator terminal 20 is an information processing device used by the evaluator of the evaluation target area. The evaluator evaluates the visit status of customer candidates in the evaluation target area. Evaluators include, for example, the following persons.
・Responsible person in charge of attracting tourism in the area, person in charge of administrative agency in the area ・Responsible person in charge of attracting tourism in the area, person in charge of tourism association or tourist area development corporation (DMO) in the area ・Attracting customers Persons in charge of various facilities who consider measures for

The evaluator terminals 20 may include various computers such as personal computers, server computers (eg, web servers, application servers, database servers, or combinations thereof). In this embodiment, as the evaluator terminal 20, a personal computer that mainly communicates with the behavior analysis server 30 will be described as an example.

The behavior analysis server 30 is a server that analyzes behavior such as visit status of customers in the evaluation target area based on designation of the evaluation target area input from the evaluator terminal 20 . Specific functions of behavior analysis performed by the behavior analysis server 30 will be described later.

The customer information management server 40 is a server that manages various information about customers.
The customer information managed by the customer information management server 40 includes at least the following.
・Personal information of customers (customer information)
- Information about the customer's action history linked to the customer ID - Information about the customer's search history linked to the customer ID The details of each of these pieces of information will be described later.

The customer DB 50 is a database that manages personal information of customers. The customer DB 50 is created, for example, based on personal information at the time of user registration of the customer terminal 10 . Details of the structure of the customer DB 50 will be described later.

Action history DB60 is a database which manages a customer's action history. The behavior history DB 60 is created by continuously acquiring the location information of the customer terminal 10, which changes over time, through wireless communication between the customer terminal 10 and GPS satellites (positioning device 81). Details of the structure of the action history DB 60 will be described later.

The search history DB 70 is a database that manages customer search histories. The search history DB 70 is created by acquiring information related to search queries input to the customer terminal 10 in chronological order when the customer searches using the search engine 100 (see FIG. 10) via the network 80. be done. Details of the structure of the search history DB 70 will be described later. It should be noted that the search by the user is not limited to the search using the customer terminal 10 whose position information is measured. If the terminal is associated with the customer ID, the search results performed using other terminals are also included in the search history. It is accumulated in DB70.

(1-1) Configuration of Customer Terminal 10 The hardware configuration of the customer terminal 10 will be described.
As shown in FIG. 1, the customer terminal 10 includes a storage device 11, a processor 12, an input/output interface 13, and a communication interface . Customer terminal 10 is connectable to at least one of input device 15 and output device 16 .

Storage device 11 is configured to store programs and data. The storage device 11 is, for example, a combination of ROM (Read Only Memory), RAM (Random Access Memory), and storage (eg, flash memory or hard disk).

Programs include, for example, the following programs.
・OS (Operating System) program ・Application (for example, web browser) program that executes information processing

The data includes, for example, the following data.
・Databases referenced in information processing ・Data obtained by executing information processing (that is, execution results of information processing)

The processor 12 is configured to implement the functions of the customer terminal 10 by activating programs stored in the storage device 11 . Processor 12 is an example of a computer.

Input/output interface 13 acquires signals (e.g., customer instructions, sensing signals, or combinations thereof) from input device 15 and outputs signals (e.g., image signals, audio signals, or combinations thereof) to output device 16. ).

The input device 15 is, for example, a keyboard, pointing device, touch panel, physical buttons, sensors (eg, camera, vital sensor, or combination thereof), or combinations thereof.

Output device 16 is, for example, a display, speakers, a printer, or a combination thereof.

The communication interface 14 is configured to control communications between the customer terminal 10 and external devices.

(1-2) Hardware Configuration of Evaluator Terminal 20 The hardware configuration of the evaluator terminal 20 will be described. FIG. 2 is a block diagram showing the configuration of the evaluator terminal 20 of this embodiment.

As shown in FIG. 2 , the evaluator terminal 20 includes a storage device 21 , a processor 22 , an input/output interface 23 and a communication interface 24 . The evaluator terminal 20 can be connected to at least one of an input device 25 and an output device 26 .

Storage device 21 is configured to store programs and data. Storage device 21 is, for example, a combination of ROM, RAM, and storage (eg, flash memory or hard disk).

Programs include, for example, the following programs.
・OS program ・Application program that executes information processing

The data includes, for example, the following data.
・Databases referenced in information processing ・Execution results of information processing

The processor 22 is configured to realize the functions of the evaluator terminal 20 by activating the programs stored in the storage device 21 . Processor 22 is an example of a computer.

The input/output interface 23 acquires signals (e.g., customer instructions, sensing signals, or combinations thereof) from the input device 25 and outputs signals (e.g., image signals, audio signals, or combinations thereof) to the output device 26. ).

Input device 25 is, for example, a keyboard, pointing device, touch panel, sensor, or a combination thereof.

Output device 26 is, for example, a display, speakers, or a combination thereof.

The communication interface 24 is configured to control communication between the evaluator terminal 20 and external devices.

(1-3) Hardware Configuration of Behavior Analysis Server 30 The hardware configuration of the behavior analysis server 30 will be described. FIG. 3 is a block diagram showing the hardware configuration of the behavior analysis server 30 of this embodiment.

As shown in FIG. 3 , behavior analysis server 30 includes storage device 31 , processor 32 , and communication interface 34 .

Storage device 31 is configured to store programs and data. Storage device 31 is, for example, a combination of ROM, RAM, and storage (eg, flash memory or hard disk).

Programs include, for example, the following programs.
・OS program ・Application program that executes information processing

The data includes, for example, the following data.
・Databases referenced in information processing ・Execution results of information processing

The processor 32 is configured to implement the functions of the behavior analysis server 30 by activating programs stored in the storage device 31 . Processor 32 is an example of a computer.

Communication interface 34 is configured to control communication between behavior analysis server 30 and external devices.

(1-4) Functional Configuration of Behavior Analysis Server 30 Next, the functional configuration of the behavior analysis server 30 will be described. FIG. 4 is a block diagram showing the functional configuration of the behavior analysis server 30. As shown in FIG.
As shown in FIG. 4 , the behavior analysis server 30 functions as a communication section 301 , a storage section 302 and a control section 303 .

(1-4-1) Functions of Communication Unit 301 and Storage Unit 302 The communication unit 301 performs processing for the behavior analysis server 30 to communicate with an external device.

Storage unit 302 stores data and programs used by behavior analysis server 30 . The storage unit 302 stores, for example, the following data.
・Mother set data 3021
・Non-visit customer data 3022
・ Visiting customer data 3023
・Total data 3024
・Reference value data 3025
・Evaluation result data 3026

The population data 3021 is information containing the ratio of each area actually visited during the evaluation period among the user groups who searched for the evaluation target area extracted from the past search history accumulated in the search history DB 70. be. In other words, the mother set data is information obtained by classifying users who are customer candidates for each area they actually visited during the evaluation period. The mother set data 3021 includes the ratio of each area visited by the user group who searched for the evaluation target area. The following information is associated with the mother set data 3021 .
・User ID related to the relevant user
・Customer information related to the relevant user ・Action log related to the relevant user ・Search log related to the relevant user

The non-visiting customer data 3022 refers to data related to. refers to information about customers who have not visited the area to be evaluated in the behavior history of the customers among the customers included in the population data 3021 .
The non-visiting customer data 3022 includes the user ID of the corresponding user. The non-visiting customer data 3022 may include the following information along with the user ID of the corresponding user.
・Customer information related to the relevant user ・Action log related to the relevant user ・Search log related to the relevant user

Visiting customer data 3023 refers to data relating to visiting customers. A visiting customer refers to a customer who has visited an area to be evaluated in the customer's behavior history among the customers included in the population data 3021 .
The visiting customer data 3023 includes the user ID of the corresponding user. The visiting customer data 3023 may include the following information together with the user ID of the corresponding user.
・Customer information related to the relevant user ・Action log related to the relevant user ・Search log related to the relevant user

Aggregated data 3024 refers to aggregated data regarding non-visiting customers or visiting customers. A specific example of the aggregated data 3024 will be described later as an output image.

The reference value data 3025 indicates an index (reference value) that serves as a reference for evaluating the evaluation target area. A specific example of the reference value data 3025 will be described later.

The evaluation result data 3026 refers to data regarding the result of evaluating the evaluation target area using the reference value. A specific example of the evaluation result data 3026 will be described later as an output image.

(1-4-2) Functions of Control Unit 303 The control unit 303 shown in FIG. 3 functions as various functional modules by the processor 32 of the behavior analysis server 30 performing processing according to programs.
Various functional modules include a transmission/reception control module 3030 , an extraction module 3032 , an aggregation module 3033 , a classification module 3034 , a reference value setting module 3035 , an evaluation module 3036 and an output module 3037 .

<Functions of transmission/reception control module 3030>
The transmission/reception control module 3030 controls processing for the behavior analysis server 30 to transmit and receive signals from an external device according to a communication protocol.

<Function of determination module 3031>
The determination module 3031 designates the visiting area and the input evaluation target area for the population data 3021, and determines whether each user group corresponding to the visiting area is a visiting customer or a non-visiting customer. Determine whether or not At this time, the determination module confirms whether or not each user group that constitutes the population data 3021 has visited the evaluation area. judge. On the other hand, when the user group to be determined has not visited the evaluation area, the user group is determined as non-visiting customers.

In addition, the determination module 3031 confirms the user's place of residence and work place, and if the evaluation area is located within a predetermined range from the user's place of residence, in the evaluation, the user is included in the population data. Even if the customer is included, it may be excluded from the determination of whether or not the customer is a non-visit customer. This makes it possible to exclude users who have searched for areas within their own living sphere, for example, due to the necessity of their daily life, from behavioral analysis.

In addition, the determination module 3031 refers to the search history of the non-visiting customer in the search history DB 70, and extracts the words written together with the name of the evaluation target area in the search query using, for example, natural language processing accompanied by text mining. , is extracted as a search target keyword.
For example, when the search query is "Matsumoto City + Gourmet", the determination module 3031 extracts the words "Matsumoto City" as the name of the evaluation area. Then, the determination module 3031 extracts the word "gourmet" written together with the name of the evaluation area as a search target keyword. In this case, it can be inferred that the user who used the search query performed a search to obtain gourmet information in Matsumoto City. In this way, the search target keyword is information that can specify the target concept that is the target of the customer's search. Note that the extraction module 3032 may refer to the correspondence table without performing the text mining process, and extract words that match the words recorded in the correspondence table as search target keywords.

<Functions of extraction module 3032>
The extraction module 3032 removes, as noise, wordings unrelated to the activity to be analyzed from among the wordings written together with the name of the evaluation target area in the search query, and extracts the remaining wordings as the search target keywords. be able to.
For example, when the search query is "Osaka + celebrity", the extraction module 3032 extracts the word "celebrity" as the word written together with the name of the evaluation target area. Then, the extraction module 3032 determines that the extracted wording “celebrity” is a wording unrelated to tourism activities, and thereby excludes the wording as noise. This is because, in this case, it can be inferred that the purpose of the search is to obtain information on, for example, a celebrity from Osaka.

The extraction module 3032 can use a noise determination model (learned model) to determine whether the words written together with the name of the evaluation target area correspond to noise. In this case, the noise judgment model is obtained by causing the machine learning model to perform machine learning according to the model learning program based on the learning data. In the present embodiment, the noise judgment model is trained so as to output a judgment result as to whether or not the input of a specific word is noise, for example. At this time, the learning data uses text data relating to a plurality of phrases as input data, and information as to whether each of the plurality of phrases corresponds to noise as correct output data.

The noise judgment model is, for example, a parameterized synthesized function obtained by synthesizing a plurality of functions. A parameterized composite function is defined by a combination of multiple tunable functions and parameters. The noise judgment model may be any parameterized composite function that satisfies the above requirements, but is assumed to be a multi-layered network model (hereinafter referred to as a multi-layered network). A noise judgment model using a multi-layered network includes an input layer for inputting words, an output layer for outputting judgment results as to whether or not there is noise, and at least one intermediate layer provided between the input layer and the output layer. Alternatively, it has a hidden layer. The noise judgment model is assumed to be used as a program module that is part of artificial intelligence software.

As the multi-layered network according to the present embodiment, for example, a deep neural network (DNN), which is a multi-layered neural network targeted for deep learning, can be used. As the DNN, for example, a recurrent neural network (RNN) targeting time-series information or the like may be used. Also, a network with a self-attention mechanism targeting natural language may be used.

In addition, the extraction module 3032 may refer to a library of wordings that can be noise of the search target keyword to determine whether the wordings written together with the name of the evaluation target area correspond to noise. In addition, the extraction module 3032 may use text mining processing to determine whether the words written together with the name of the evaluation target area correspond to noise.

<Functions of Aggregation Module 3033>
The tallying module 3033 uses the non-visiting customer data 3022 to refer to the customer information on the non-visiting customers in the customer DB 50, and tallies the information related to the non-visiting customers. Information related to non-visiting customers aggregated by the aggregation module 3033 includes, for example, the following.
・Proportion of non-visiting customers in the mother set data 3021 ・Distribution of customer segments included in non-visiting customers ・Search queries with high search frequency among search queries including the name of the evaluation target area and their search frequency ・Evaluation area Visited area instead of In addition, these aggregated results may be aggregated for each customer segment of the user.
Here, a customer segment refers to a customer group that classifies a group of users based on any attribute managed as customer information, such as sex, age, and occupation.
Aggregation of this information is described in detail below.

Aggregation module 3033 aggregates the proportion of non-visit customers in mother set data 3021 for the evaluation target area. This makes it possible to grasp what percentage of the customers who searched during the evaluation period actually visited the evaluation target area. At this time, the ratio of non-visiting customers in the mother set data 3021 is also aggregated at the same time.

Aggregation module 3033 aggregates the distribution of customer segments included in non-visit customers. At this time, the tabulation module 3033 refers to the customer DB 50, classifies the users based on any attribute item managed as customer information, and tabulates the distribution of customers for each item. Specifically, users are classified according to items such as age group, sex, and place of residence, and the distribution of customers for each item is aggregated. At this time, an evaluation item may be set by multiplying items related to a plurality of attributes.

The aggregation module 3033 aggregates search queries with high search frequency among search queries including the name of the evaluation target area. At this time, the counting module 3033 refers to the search history of the non-visiting customer in the search history DB 70, and counts the search queries with high search frequency among the search queries including the name of the evaluation area.

The counting module 3033 refers to the action history of the non-visiting customer in the action history DB 60, and specifies the visited area instead of the evaluation area as the competitive area. Here, the competitive area refers to an area that is highly probable to compete with each other in attracting customers to visit the area because it has some characteristics in common with the area to be evaluated as a tourist destination.
For example, as areas that are geographically relatively close to each other and have similar climates, natural resources, and accessibility, Nagano City, Shiojiri City, and Azumino City are all expected to compete with Matsumoto City. be.
In addition, the counting module 3033 refers to the past search history, and among the search contents performed at the same time, identifies areas searched with the same search purpose keyword as the area to be evaluated as competitive areas. may

On the other hand, areas such as Sapporo, which are located far from Matsumoto, have different climates, natural resources, and accessibility, so they are less likely to compete with Matsumoto. Conceivable. Therefore, the tallying module 3033 may exclude, from among the areas visited by non-visiting customers in place of the evaluation target area, areas separated from the evaluation target area by a predetermined distance or more from the competitive area candidates. Exclusion from such competitive area candidates can be arbitrarily executed by referring to the distance from the evaluation area as well as accessibility and differences in natural resources.

In addition, using the visiting customer data 3023, the tabulation module 3033 can also refer to customer information about visiting customers in the customer DB 50 and tabulate information related to visiting customers. Information related to visiting customers aggregated by the aggregation module 3033 includes, for example, the following. Specifically, since it is the same as the tabulation for the non-visiting customers described above, the explanation thereof will be omitted.
・Percentage of visiting customers in the mother set data 3021 ・Distribution of customer segments included in visiting customers ・Search queries with high search frequency among search queries including the name of the evaluation target area and their search frequency Results may be aggregated by the user's customer segment.

In addition, the aggregation module 3033 can aggregate the visited areas together with the visits to the evaluation areas as information related to visiting customers. At this time, the tallying module 3033 refers to the behavior history of the visiting customer in the movement history DB, and identifies the area visited together with the visit to the evaluation target area as a cooperative area. Here, the coordinated area refers to an area that is close to the evaluation target area and that each has its own unique tourist attraction characteristics, so that there is a high probability that customers will want to tour together. For example, Matsumoto City, which is adjacent to each other and has a castle, and Azumino City, which is blessed with mountainous scenery, can be a competitive area as described above, but at the same time it is presumed that they can also be a cooperative area. In addition, the determination of whether or not they visited together is based on the action history indicated by a series of trip chains, regardless of whether it is for multiple days or a single day, when the place of residence is the starting point and the ending point in the trip chain that shows the movement route. This can be determined by whether you have visited both areas in the

The aggregation module 3033 can aggregate information related to non-visit customers for each search target keyword and for each keyword category described later. The specific contents of the processing will be described later.
The tabulation module 3033 also refers to the search histories of visiting customers in the search history DB 70, and tabulates keyword categories to which search purpose keywords that frequently appear in search queries correspond as strengths of the evaluation target area. The specific contents of the processing will be described later.

<Functions of Classification Module 3034>
The classification module 3034 classifies each of the plurality of evaluation areas into a plurality of area categories set based on at least one of the evaluation index indicating the scale of the city and the tourist spot characteristics. Evaluation indicators that indicate the size of a city include, for example, the following items.
・Daytime population ・Resident population within the time zone ・Indices showing size of metropolitan area and living area

The tourist spot characteristics include tourism categories known as tourism objectives for the evaluation area. Sightseeing categories include, for example, the following items.
・Accommodation: Activities for the purpose of staying at lodging facilities ・Eating and drinking: Activities for the purpose of receiving food and drinks ・Summer retreat: Activities for the purpose of visiting summer resorts ・Hot springs: Visiting hot springs Activities aimed at ・Sea bathing: Activities aimed at bathing in the sea ・Amusement: Visiting downtown areas ・Outdoor activities: Activities aimed at outdoor activities such as camping and mountain climbing ・Activities: Marine sports, athletics, and other exercises・History exploration: Activities aimed at visiting historical and cultural assets such as castles, historic sites, temples and shrines ・Amusement: Activities aimed at visiting theme parks, amusement parks, and leisure facilities ・Curture: Visiting museums and art galleries Activities for the purpose of visits/Other: Activities for the purpose of acts that do not fall under any of the above (e.g. charity activities, etc.)

Classification module 3034 can classify evaluation areas into area categories using area classification models (learned models). In this case, the area classification model is obtained by having the machine learning model perform machine learning according to the model learning program based on the learning data. In this embodiment, the area classification model is trained so that, for example, in response to input of words indicating the name of a specific evaluation area, a corresponding area category is output. At this time, the learning data uses the names of the evaluation areas as input data, and the names of the area category divisions to which the names of the plurality of evaluation areas correspond as correct output data.

The area classification model is, for example, a parameterized composite function composed of multiple functions. A parameterized composite function is defined by a combination of multiple tunable functions and parameters. The area classification model can be any parameterized composite function that satisfies the above requirements, but is assumed to be a multi-layered network model (hereinafter referred to as a multi-layered network). An area classification model using a multi-layered network has an input layer for inputting words, an output layer for outputting the names of corresponding area category divisions, and at least one intermediate layer provided between the input layer and the output layer. Alternatively, it has a hidden layer. The area classification model is assumed to be used as a program module that is part of artificial intelligence software.

As the multi-layered network according to the present embodiment, for example, a deep neural network (DNN), which is a multi-layered neural network targeted for deep learning, can be used. As the DNN, for example, a recurrent neural network (RNN) targeting time-series information or the like may be used. Also, a network with a self-attention mechanism targeting natural language may be used.

Also, the classification module 3034 may refer to a classification table in which area categories are defined for each area to classify the evaluation areas.

Also, the classification module 3034 classifies the search target keyword extracted by the extraction module 3032 into a category (keyword category) corresponding to a division corresponding to the meaning thereof. A keyword category is a classification division in which a concept including a search target keyword is converted into a higher-level concept as the purpose of visiting the evaluation area. Although the name of the keyword category division can be set arbitrarily, for example, the sightseeing division described above can be used.

The classification module 3034 can classify the extracted search target keywords into keyword categories using a keyword classification model (learned model). In this case, the keyword classification model is obtained by causing the machine learning model to perform machine learning according to the model learning program based on the learning data. In this embodiment, the keyword classification model is learned so as to output a corresponding keyword category, for example, in response to input of phrases indicating a specific search target keyword. At this time, the training data uses text data relating to a plurality of phrases as input data, and the names of the keyword categories to which each of the plurality of phrases corresponds as correct output data.

A keyword classification model is, for example, a parameterized composite function composed of multiple functions. A parameterized composite function is defined by a combination of multiple tunable functions and parameters. The keyword classification model can be any parameterized composite function that satisfies the above requirements, but is assumed to be a multi-layered network model (hereinafter referred to as a multi-layered network). A keyword classification model using a multi-layered network includes an input layer for inputting text, an output layer for outputting the name of the category of the corresponding keyword category, and at least one intermediate layer provided between the input layer and the output layer. Alternatively, it has a hidden layer. The keyword classification model is assumed to be used as a program module that is part of artificial intelligence software.

As the multi-layered network according to the present embodiment, for example, a deep neural network (DNN), which is a multi-layered neural network targeted for deep learning, can be used. As the DNN, for example, a recurrent neural network (RNN) targeting time-series information or the like may be used. Also, a network with a self-attention mechanism targeting natural language may be used.

The classification module 3034 may also refer to a library of keyword categories to determine whether the wording of the search target keyword corresponds to noise. The classification module 3034 may also use a text mining process to determine which keyword category the wording of the search target keyword falls under. Also, the classification module 3034 may refer to a classification table in which keyword categories are defined for each keyword, and classify the keyword categories.

<Function of reference value setting module 3035>
The reference value setting module 3035 sets a reference value indicating the degree of non-visiting customers for each of a plurality of area categories. A reference value is set by relatively evaluating the amount and ratio of non-visiting customers in a plurality of evaluation areas included in each area category. Specific processing of the reference value setting module 3035 will be described later.

<Functions of Evaluation Module 3036>
The evaluation module 3036 performs evaluation based on a reference value for the evaluation area to be evaluated. Specific processing of the evaluation module 3036 will be described later.

<Function of output module 3037>
The output module 3037 generates output image data from the processing results of the functional modules described above, and outputs the output image data to the evaluator terminal 20 . A specific output image will be described later.

(1-5) Hardware Configuration of Customer Information Management Server 40 The hardware configuration of the customer information management server 40 will be described. FIG. 5 is a block diagram showing the hardware configuration of the customer information management server 40 of this embodiment.

As shown in FIG. 5, the customer information management server 40 includes a storage device 41, a processor 42, and a communication interface 44.

Storage device 41 is configured to store programs and data. Storage device 41 is, for example, a combination of ROM, RAM, and storage (eg, flash memory or hard disk).

Programs include, for example, the following programs.
・OS program ・Application program that executes information processing

The data includes, for example, the following data.
- Databases referenced in information processing - Execution results of information processing (for example, mother set data 3021)

The processor 42 is configured to implement the functions of the customer information management server 40 by activating programs stored in the storage device 41 . Processor 42 is an example of a computer.

The communication interface 44 is configured to control communication between the customer information management server 40 and external devices.

(2) Structure of database The structure of each database used in this embodiment will be described. The structure of each database described below is merely an example, and can be arbitrarily changed.

(2-1) Customer database 50
FIG. 6 is a diagram showing a specific example of the customer DB 50. As shown in FIG.
As shown in FIG. 6, the customer DB 50 includes an item "user ID", an item "name", an item "sex", an item "age group", an item "residence", an item "occupation", It has an item "place of work" and an item "terminal ID".

The item “user ID” stores user identification information that can identify individual users.

The item "name" stores information about the name of the user corresponding to the user ID.

The item “gender” stores information about the gender of the user corresponding to the user ID.

The item "age group" stores information about the age group of the user corresponding to the user ID.

The item "place of residence" stores information about the place of residence of the user corresponding to the user ID.

The item "occupation" stores information about the occupation of the user corresponding to the user ID.

The item "place of work" stores information about the place of work of the user corresponding to the user ID.

The item "terminal ID" stores the identification information of the customer terminal 10 used by the user corresponding to the user ID.
Note that the structure of the customer DB 50 shown in FIG. 6 is merely an example, and other items may be managed.

(2-2) Action history database 60
FIG. 7 is a diagram showing a specific example of the action history DB 60. As shown in FIG.
As shown in FIG. 7, the action history DB 60 includes an item "action log ID", an item "user ID", an item "location information", and an item "measurement date and time".

The item “action log ID” stores identification information that can identify each action log.

The item "user ID" stores the user ID of the user who uses the customer terminal 10 from which the action log corresponding to the action log ID was transmitted.

The item “location information” stores location information (coordinate information) included in the action log corresponding to the action log ID.

The item "measurement date and time" stores information on the date and time when the action log corresponding to the action log ID was measured.
Note that the structure of the action history DB 60 shown in FIG. 7 is merely an example, and other items may be managed.

(2-3) Search history database 70
FIG. 8 is a diagram showing a specific example of the search history DB 70. As shown in FIG.
As shown in FIG. 8, the search history DB 70 includes the item "search log ID", the item "user ID", the item "search query", and the item "search date and time".

The item “search log ID” stores identification information that can identify individual search logs.

The item "user ID" stores the user ID of the user who uses the terminal that executed the search log corresponding to the search log ID.

The item “search query” stores text data indicating a search query input in the search log corresponding to the search log ID.

The item "search date and time" stores information on the date and time when the search log corresponding to the search log ID was executed.
Note that the structure of the search history DB 70 shown in FIG. 8 is merely an example, and other items may be managed.

(3) Outline of system 1 according to this embodiment An outline of the system 1 according to this embodiment will be described.
FIG. 9 is a diagram illustrating the first half of the outline of the system 1 in this embodiment.

As shown in FIG. 9, the system 1 uses mobile communication between the customer terminal 10 and a positioning device (GPS antenna) 81 to estimate and acquire the location information of the customer terminal 10 along time series in a wide area. , the action history DB 60 is created. The action history DB 60 is operated, for example, by a management company of application software installed in the terminal 10 that acquires position information. The position information acquired by the positioning device 81 is stored in the action history DB 60 in association with the user ID together with the time information.

The action logs accumulated in the action history DB 60 comprehend chronological changes, thereby forming trip data indicating a travel route from the starting point to the ending point. By confirming the trip data, the user carrying the customer terminal 10 can grasp the place visited, the date and time, the moving route, the staying time, and the like. That is, a set of action logs becomes an action history.

FIG. 10 is a diagram illustrating the second half of the outline of the system 1 in this embodiment.
As shown in FIG. 10, for example, a user who wants to travel somewhere searches for tourist attractions of interest in an area of interest to the user via the customer terminal 10 . At this time, the search engine 100 that has received the search query from the customer terminal 10 outputs search information in response to the input search query. The search engine 100 then records a search log in the search history DB 70 . The search history DB 70 is operated by, for example, the management company of the search engine 100 . In the search history DB 70, for example, text data of search queries when a search is performed by the search engine 100 using the customer terminal 10 is stored in association with the user ID together with the search date and time.

Then, in the system 1, the user who searched the evaluation target area is regarded as a customer candidate for the area. Among the customer candidates, users who have actually visited are identified as visiting customers, and users who have not visited among the customer candidates are identified as non-visiting customers.

A series of these processes will be described in detail below. It is assumed that the customer DB 50, the action history DB 60, and the search history DB 70 have already been constructed when step S201 of FIG. 11 is performed.

(4) Processing Flow Specific processing of the system 1 will be described below based on each flow.

(4-1) First Example of Processing by System 1 FIG. 11 is a diagram illustrating a first example of processing by the system 1. FIG. This process is a process of aggregating information about non-visit customers.

As shown in FIG. 11, in the system 1, the evaluator first performs an input operation of evaluation conditions using the evaluator terminal 20 (step S201). Evaluation conditions include an evaluation area and an evaluation period. If the evaluation period has already been set, only the evaluation area may be entered.
Specifically, the evaluator inputs an area name for designating the area to be evaluated.

After step S201, the behavior analysis server 30 receives input of evaluation conditions (step S301).
Specifically, the transmission/reception control module 3030 of the behavior analysis server 30 receives the input information regarding the evaluation area. At this time, if the evaluation period is specified, that information is also accepted at the same time.

After step S301, the customer information management server 40 sets the evaluation area based on the information transmitted from the behavior analysis server 30 (step S400).
Specifically, the transmission/reception control module 3030 of the behavior analysis server 30 transmits the input information regarding the evaluation area to the customer information management server 40 . Customer information management server 40 accepts the transmitted information.

After step S400, the customer information management server 40 creates population data 3021 for the set evaluation area (step S401).
Specifically, the processor 42 of the customer information management server 40 refers to the search history DB 7080 and extracts a search log of searching the specified evaluation area during the set evaluation period. The processor 42 refers to the customer DB 50 from the extracted user ID that performed the search log, and identifies the user who is the subject of the search log as a customer candidate.
The processor 42 refers to the action history DB and specifies the user's visit area during the evaluation period by specifying actions outside the living area from the action history during the evaluation period of the specified user. Then, the processor 42 classifies the user group that constitutes the customer candidate by visit area, and aggregates the scale, thereby creating population data 3021 .

After step S401, the customer information management server 40 transmits the created population data 3021 to the behavior analysis server 30 (step S402).
Specifically, the processor 42 of the customer information management server 40 analyzes, as the created mother set data 3021, for example, aggregate results showing a breakdown of scale for each visited area for a group of users constituting customer candidates. Send to server 30 . At this time, the search log and action log corresponding to the user ID may be included in the mother set data 3021 .

After step S402, the behavior analysis server 30 acquires the population data 3021 (step S302).
Specifically, the transmission/reception control module 3030 of the behavior analysis server 30 acquires the collective data transmitted from the customer information management server 40 and stores it in the storage unit 302 .

After step S303, the behavior analysis server 30 determines non-visiting customers from the population data 3021 (step S303).
Specifically, the determination module 3031 of the behavior analysis server 30 compares the specified evaluation area with the visited areas visited by the user group that constitutes the customer candidate, thereby determining the number of users included in the population data 3021. Among the group, customers who have not visited the evaluation area are determined as non-visit customers. The determination module 3031 causes the storage unit 302 to store information about non-visiting customers linked to the mother set data as non-visiting customer data 3022 .

After step S303, the behavior analysis server 30 aggregates information related to non-visit customers (step S304).
Specifically, the aggregation module 3033 of the behavior analysis server 30 refers to the non-visiting customer data 3022 stored in the storage unit 302, and aggregates the following information, for example.
・Proportion of non-visiting customers in the mother set data 3021 ・Distribution of customer segments included in non-visiting customers ・Search queries with high search frequency among search queries including the name of the evaluation target area and their frequency of retrieval ・Non-visiting customers Areas Visited by Customers in Place of Evaluation Areas These aggregated results may also be aggregated for each customer segment of the user.
Aggregation module 3033 stores the aggregated information in storage unit 302 as aggregated data 3024 .

After step S304, the behavior analysis server 30 outputs the aggregated information (step S305).
Specifically, the output module 3037 of the behavior analysis server 30 creates an output image based on the aggregated data 3024 stored in the storage unit 302 and transmits it to the evaluator terminal 20 via the transmission/reception control module 3030 . A specific example of the output image will be described later.

After step S305, the evaluator terminal 20 displays an output image based on the output total data 3024 (step S202).
Specifically, the processor 22 of the evaluator terminal 20 outputs the output image output by the output module 3037 of the behavior analysis server 30 to the output device 26 of the evaluator terminal 20 . As a result, the evaluator is presented with the aggregate result of the information on the non-visiting customer.

(4-2) Second Example of Processing by System 1 FIG. 12 is a diagram illustrating a first example of processing by the system 1. FIG.
This process is a process of totaling information about visiting customers, corresponding to the previously described totaling of information about non-visiting customers. In this process, the processes from step S201 to step S302 are the same as those of the first example shown in FIG. 11, so repeated descriptions will be omitted.

After step S302 shown in FIG. 12, the behavior analysis server 30 determines visiting customers for the user group included in the mother set data 3021 (step S303B).
Specifically, the determination module 3031 of the behavior analysis server 30 compares the specified evaluation area with the visited areas visited by the user group that constitutes the customer candidates, thereby determining the number of customers included in the mother set data 3021. Among them, a customer who visited the evaluation area during the evaluation period is determined as a visiting customer. The determination module 3031 causes the storage unit 302 to store information about visiting customers linked to the mother set data as visiting customer data 3023 .

After step S303B, behavior analysis server 30 aggregates information related to visiting customers (step S304B).
Specifically, the aggregation module 3033 of the behavior analysis server 30 refers to the visiting customer data 3023 stored in the storage unit 302, and aggregates the following information, for example.
・Percentage of visiting customers in mother set data 3021 ・Distribution of customer segments included in visiting customers ・Search queries with high search frequency among search queries including the name of the evaluation target area and their search frequency ・Combined with the evaluation area Also, these aggregated results may be aggregated for each customer segment of the user.
The tabulation module 3033 stores the tabulated information in the storage unit 302 as tabulated data 3024 .

After step S304B, processing for displaying the totaled information on the evaluator terminal 20 is performed in the same manner as in steps S305 and 202 shown in FIG. Since these processes are the same as those of the first example shown in FIG. 11, repeated description is omitted.

12 may be performed at the same time as the processing for aggregating information about non-visiting customers described with reference to FIG. That is, the determination of visiting customers and non-visiting customers can be evaluated simultaneously. FIG. 13 is a diagram for explaining an example of determination of visiting customers and non-visiting customers by the system 1. In FIG.

As shown in FIG. 13, for example, when the evaluation area is Matsumoto City. Determination of the customer type is as follows with respect to the mother set of users (customer candidates) who performed a search with the search query "Matsumoto sightseeing" within the evaluation period.
・Users who visited Matsumoto during the evaluation period: visiting customers ・Users who visited places other than Matsumoto: lost customers A behavioral customer constitutes a non-visiting customer who is a potential customer.
It should be noted that a non-visiting customer may be determined as a user with low demand for travel and not a non-visiting customer (not subject to evaluation).

Next, when evaluating the areas actually visited, if it can be confirmed that the visiting customer has visited (toured) Nagano City, Azumino City, or Takayama City along with the visit to Matsumoto City, these areas is determined to be a cooperative area.

On the other hand, if it is confirmed that the lost customer visited Nagano City, Azumino City, Kusatsu City, Kanazawa City, Toyama City, Sapporo City during the evaluation period, Nagano City, Azumino City, Kusatsu City, Kanazawa City and Toyama City are determined to be competing areas. Here, since Sapporo City is a remote area from Matsumoto City, it is excluded from the competitive area.

That is, Nagano City and Azumino City are determined to be both a cooperative area and a competitive area. In this case, it is possible to judge whether Nagano City and Azumino City have a strong aspect as a cooperative area or a strong aspect as a competitive area from the ratio of each visit pattern shown in the whole.

(4-3) Keyword Analysis Processing by System 1 FIG. 14 is a diagram for explaining keyword analysis processing by the system 1 .
This process refers to the search history DB 70, analyzes the words included in the search query, and uses them for subsequent behavior analysis.

As shown in FIG. 14, in the keyword analysis process, the behavior analysis server 30 extracts a search target keyword (step S311).
Specifically, the extraction module 3032 of the behavior analysis server 30 refers to the search history DB 70 and extracts a search target keyword from search queries included in the search histories of visiting and non-visiting customers. At this time, the extraction module 3032 extracts the words written together with the name of the evaluation area from the words included in the search query, and then uses, for example, a noise determination model to extract the words written together with the name of the evaluation area. Determine whether or not

A specific example of extracting a search target keyword using "Matsumoto City" as the evaluation area will be shown.
Example 1) Search query: "Matsumoto Castle" ⇒Extract "castle" as a search target keyword In this case, "castle" is extracted as a word written together with the name of the evaluation area. This search query is considered to be for the purpose of obtaining information (sightseeing purpose) on Matsumoto Castle, which is a tourist resource in Matsumoto City.

・Example 2) Search query: "Matsumoto Yakiniku" ⇒Extract "Yakiniku" as a search target keyword In this case, "Yakiniku" is extracted as a word written together with the name of the evaluation area. There is a plan to eat in Matsumoto City, and it is considered to be a search query for the purpose of acquiring information on restaurants (sightseeing purposes in a broad sense).

・Example 3) Search query: “Matsumoto cheap hotel” ⇒ Extract cheap hotel as a search target keyword In this case, “cheap hotel” is extracted as a word written together with the name of the evaluation area. It is thought that the search was aimed at obtaining information on accommodation facilities when visiting Matsumoto City (for tourism purposes in a broad sense).

·Example 4) Search query: "Matsumoto Age" ⇒ Remove the word written together as noise In this case, "age" is extracted as the word written together with the name of the evaluation area. In this case, it is considered that the search is not for the purpose of sightseeing, but for the purpose of obtaining the personal information of a celebrity with the surname "Matsumoto."

After step S312, the behavior analysis module classifies the extracted search purpose keywords into keyword categories (step S312).
Specifically, the classification module 3034 of the behavior analysis module classifies the extracted search target keyword into preset keyword categories (for example, lodging, eating and drinking, summer vacation, hot springs, sea bathing, entertainment, outdoors, activities, historical exploration, entertainment, curatorial, and others).

The keywords in Examples 1 to 3 above correspond to the following.
・Example 1) Keyword for search purpose: “Castle” ⇒ Classify into keyword category “historical visit” ・Example 2) Keyword for search purpose: “Yakiniku” ⇒ Classify into keyword category “restaurant” ・Example 3) Keyword for search purpose: “ Cheap hotels” ⇒ classified into the keyword category “accommodation”

After step S312, the behavior analysis module aggregates information on non-visit customers and visiting customers for each search target keyword (step S313).
Specifically, the tabulation module 3033 of the behavior analysis module tabulates the ratio of non-visit customers and visiting customers in the evaluation area (for example, Matsumoto City) as follows.
・Overall: 30% of visiting customers / 70% of non-visiting customers ・Search target keyword “castle”: 50% of visiting customers / 50% of non-visiting customers I can confirm. In other words, it can be inferred that Matsumoto City has high latent demand for customers who are interested in visiting castles.
After that, the behavior analysis server 30 outputs the tabulation result to the evaluator terminal 20, as in FIGS. 11 and 12. FIG.

After step S313, the behavior analysis module aggregates information on non-visit customers and visiting customers for each keyword category (step S314).
Specifically, the tabulation module 3033 of the behavior analysis module tabulates the ratio of non-visit customers and visiting customers in the evaluation area (for example, Matsumoto City) as follows.
・Overall: Visiting customers 30% / Non-visiting customers 70% ・Keyword category “History visit”: Visiting customers 50% / Non-visiting customers 50% I can confirm. In other words, it can be inferred that Matsumoto City has high latent demand for customers who are interested in exploring history.
After that, the behavior analysis server 30 outputs the tabulation result to the evaluator terminal 20, as in FIGS. 11 and 12. FIG.

(4-4) Area Analysis Processing by System 1 FIG. 15 is a diagram for explaining area analysis processing by the system 1 .
This process is a process of classifying the evaluation areas in advance and relatively evaluating the evaluation areas included in the same category.

As shown in FIG. 15, in the area analysis process, the behavior analysis server 30 classifies the evaluation area into area categories (step S311).
Specifically, the classification module 3034 of the behavior analysis server 30 classifies each of the plurality of evaluation areas into area categories set based on one of the following indices.
·Evaluation index indicating scale of city ·Tourist spot characteristics That is, the classification module 3034 classifies the evaluation area based on its characteristics as a tourist spot and the scale of the city. For example, Matsumoto City has a population of approximately 240,000, making it a medium-sized city (100,000 or more but less than 500,000). Matsumoto Castle, a tourist resource, is well-known, and is popular for historical exploration. Therefore, as an area category, it can be classified into the division of "medium city/historical exploration".

After step S311, the behavior analysis server 30 sets a reference value indicating the degree of non-visiting customers for each set area category. Specifically, the reference value setting module 3035 in the behavior analysis server 30 relatively evaluates the amount and ratio of non-visit customers in a plurality of evaluation areas included in each area category, thereby setting an appropriate reference value.

For example, the reference value setting module 3035 refers to the amount and ratio of non-visit customers in each of a plurality of evaluation areas included in the area category "medium city/historical visit", Calculate the volume and percentage of typical non-visit customers.

Then, the reference value setting module 3035 sets each calculated value as a reference value in the area category. Note that the standard value setting module 3035 may calculate the amount and percentage of standard non-visit customers by various mathematical processes such as arithmetic average, weighted average, geometric average, or harmonic average. Also, the reference value setting module 3035 may calculate the reference value using a function for calculating the reference value with the amount and ratio of non-visit customers as variables.
That is, the reference value may be a numerical value of the same dimension as the amount and ratio of non-visiting customers, or may be a dimensionless value calculated from the amount and ratio of non-visiting customers.

After step S322, the behavior analysis server 30 evaluates the evaluation area based on the set reference value (step S323).
Specifically, the evaluation module 3036 of the behavior analysis server 30 compares the reference value with the amount and ratio of non-visit customers in the evaluation area to calculate relative performance within the same area category. After that, the behavior analysis server 30 outputs the evaluation result to the evaluator terminal 20 as in FIGS. 11 and 12 .

(5) Output Image Below, a specific output image by the system 1 will be described based on each screen example.

(5-1) First Example of Output Image FIG. 16 is a diagram showing a first example of an output image of the system 1. FIG. Among them, FIG. 16A is a diagram showing the ratio of visiting customers and non-visiting customers. By showing the ratio of visiting customers and non-visiting customers as shown in FIG. 16A, it is possible to suggest to the evaluator the amount of non-visiting customers in the evaluation area. The tallying module may tally differences in the ratio and number of visiting customers and non-visiting customers, and output the result.

Next, FIG. 16B is a diagram showing the customer distribution of visiting customers and non-visiting customers. In this distribution, the distribution for each attribute can be confirmed by selecting the attribute item located at the top. By confirming the distribution of such customer segments, it becomes easier for the evaluator to grasp the customer segments that should be targeted in acquiring non-visit customers.

Next, FIG. 16C is a diagram showing the ranking of search target keywords of visiting customers and non-visiting customers. In this way, by displaying the ranking of the search target keywords of visiting customers, it suggests to the evaluator the items that are of interest in the evaluation area (strengths as a tourist destination, such as famous tourist resources). be able to.
In addition, by displaying the ranking of the search purpose keywords of non-visiting customers, it is possible to suggest to the evaluator items that may be related to the reasons why the non-visiting customers did not visit.

Next, FIG. 16D is a diagram showing the ranking of visiting customers and non-visiting customers. In this way, by displaying the ranking of excursion destinations of visiting customers, it is possible to suggest to the evaluator an area that can be an effective partner in tourism promotion measures.
In addition, by displaying the ranking of the destinations visited by non-visit customers, it is possible to suggest to the evaluator areas that should be considered as competitors in measures to attract tourism.

(5-2) Second Example of Output Image FIG. 17 is a diagram showing a second example of an output image of the system 1. FIG. This output image displays detailed information about the visiting customer's excursion destination. Specifically, whether the visiting customers visited only the evaluation area (Matsumoto City in this case) or whether they also visited other areas is displayed together with the percentage. As a result, it is possible to quantitatively suggest to the evaluator whether each area other than the evaluation area has a strong aspect as a cooperative area or a strong aspect as a competitive area.

(5-3) Third Example of Output Image FIG. 18 is a diagram showing a third example of an output image of the system 1. FIG. This output image displays the evaluation results using area analysis. Specifically, the pie chart on the left shows the percentage of visiting customers and non-visiting customers in the evaluation area. The score in the area category to which the evaluation area belongs is displayed as a bar graph. On the right side, the ratio of visiting customers and non-visiting customers in other areas is displayed for reference. The figure below shows the reference value for evaluation in the corresponding area category and the actual score within the same category. By checking this figure, it is possible to check the positioning of the evaluation areas within the same area category.
Note that the group size index related to the degree of interest is an index that indicates the size of the search group in each evaluation area.

(5-4) Fourth Example of Output Image FIG. 19 is a diagram showing a fourth example of an output image of the system 1. FIG. This output image displays the aggregated results using keyword analysis. Specifically, in the table, user behavior analysis is performed by evaluating multiple search histories.

In this tallying process, user behavior analysis is performed according to the following evaluation method.
- Evaluation procedure: For each search query in search 2 performed after search 1, the evaluation target user evaluates the proportion of the user group corresponding to the following.
・Definition of user group User group A: User group who performed search 1 with the search query “Matsumoto (evaluation area name) sightseeing” User group B: After search 1, “Matsumoto (evaluation area name) keyword X User group C: User group B: User group AB: User group A searched by keyword X User group that does not visit User group BC: User group A who performed a search using keyword X but became a non-visiting customer Analysis result: Aggregate table of the ratio of each user group described above Output: Radar chart of distribution for each keyword category for user group AB, user group BC, and user group C

In this way, by checking the summary table showing the distribution of each user group for each search query, the evaluator can quantitatively grasp the ratio of each user group for each search target keyword. In this table, the search time lag indicates the average number of days from when Search 1 is performed until when Search 2 is performed. Each search query is assigned a keyword category to which the search target keyword included in the search query corresponds.

Also, by checking the radar chart 1, it is possible to intuitively grasp the distribution of each keyword category for each user group, and to evaluate the visit status of the evaluation area from various aspects.
For example, by checking the radar chart 1, it can be confirmed that the search target keyword belonging to the keyword category 4 has not been searched. That is, it can be inferred that the evaluation area is a sightseeing section with a low customer attraction power.

Further, for example, by checking the radar chart 2, it can be confirmed that there are many users who have searched for a search target keyword belonging to the keyword category 3 but become non-visiting customers. That is, in the evaluation area, it can be inferred that the tourism category corresponding to keyword category 3 has a large increase in the number of visitors due to tourism attraction.

Further, for example, by checking the radar chart 3, it can be confirmed that there are many users who searched for the search target keyword belonging to the keyword category 1 and became visiting customers. That is, in the evaluation area, it can be inferred that the tourism category corresponding to keyword category 1 has a high customer attraction power in the evaluation area and is a sightseeing category that is strong in attracting tourism.

(6) Summary As described above, according to the system 1, the behavior analysis server 30 detects customers who have not visited the evaluation area during the evaluation period among the customer candidates who have searched the evaluation area in the past. Identifies you as a customer and aggregates information related to non-visit customers.
For this reason, it is possible to collect information not only on customers who actually visited, but also on customers who were once interested but did not visit, contributing to attracting customers to visit the evaluation target area. information can be provided.

Further, in the system 1, the percentage of non-visiting customers in the mother set data 3021, which are customer candidates who have searched the evaluation area in the past, is tallied. Therefore, it is possible to quantitatively provide the evaluator with information on how much latent demand there is in the evaluation area.

In addition, the system 1 classifies customer segments included in non-visit customers based on any item managed as customer information, and aggregates the distribution of customers for each item. Therefore, it is possible to provide the evaluator with a high resolution what kind of attribute a customer can be a non-visiting customer.

In addition, the system 1 refers to the search history of non-visiting customers, and totals search queries with high search frequency among the search queries including the name of the evaluation area. Therefore, it is possible to provide the evaluator with information that can suggest the reason why the non-visiting customer did not visit.

In addition, the system 1 refers to the action history of the non-visiting customer, and identifies the visited area instead of the evaluation area as the competitive area. Therefore, in the evaluation area, it is possible to suggest to the evaluator an area that is highly likely to compete with each other in order to attract customers to sightseeing. This makes it easier for the evaluator to see areas for improvement when considering measures to attract tourism, and makes it easier for them to compare areas where they are competing, contributing to the consideration of effective measures.

Further, in the system 1, among the areas visited in place of the evaluation area, areas separated from the evaluation area by a predetermined distance or more are excluded from the competition area candidates. Therefore, by preliminarily excluding areas that can become noise, it is possible to suggest areas with high validity as areas that are highly likely to compete with each other in order to attract customers to sightseeing.

Further, the system 1 refers to the behavior history of the visiting customer, and identifies the area visited together with the visit to the evaluation area as a cooperative area. For this reason, it is possible to suggest to the evaluator areas in which there is a high probability that the customer will want to tour together. As a result, it is possible to contribute to the examination of effective measures, such as the evaluator examining measures for attraction in cooperation with the cooperative area.

Moreover, in the system 1, after extracting a search target keyword from a search query, information related to non-visiting customers is aggregated for each search target keyword. For this reason, by quantitatively suggesting the lost rate and amount for each customer group with similar search purposes, it becomes easier for the evaluator to grasp the strengths and weaknesses of the evaluation area.

Furthermore, in the search query, the system 1 removes, as noise, wordings unrelated to tourism activities among the wordings written together with the name of the evaluation area. Therefore, search histories unrelated to visits to the evaluation area can be eliminated, and accuracy of subsequent analysis using the search target keyword can be improved.

In addition, the system 1 classifies the extracted search target keywords into keyword categories corresponding to tourism classifications corresponding to their meanings, and aggregates information related to non-visit customers for each keyword category. Therefore, by quantitatively suggesting the lost rate and amount for each customer group with similar tourism purposes, it becomes easier for the evaluator to grasp the strengths and weaknesses of the evaluation area more comprehensively.

In addition, the system 1 refers to the search history of the visiting customer, and outputs the keyword category to which the search target keyword frequently appearing in the search query corresponds as the strength of the evaluation area. Therefore, in the evaluation area, it is possible to suggest to the evaluator tourism resources that have the ability to attract customers.

Further, in the system 1, the evaluation areas are classified into area categories, a reference value indicating the degree of non-visiting customers is set for each area category, and the evaluation area is evaluated based on the reference value. Therefore, it is possible to more accurately suggest to the evaluator the status of visits by customers in the evaluation area by comparing areas with similar population sizes and similar characteristics as tourist spots.

Moreover, in System 1, the daytime population, the resident population within the time zone, the size of the metropolitan zone, and the living zone are used as evaluation indexes indicating the scale of the city. Therefore, it is possible to perform an evaluation taking into consideration the scale of the evaluation area.

In system 1, well-known sightseeing classifications are used as tourist destination characteristics as the purpose of sightseeing in the evaluation area. Therefore, it is possible to perform an evaluation taking into account the characteristics of the evaluation area as a tourist spot.

Further, in system 1, the reference value is set by relatively evaluating the amount and ratio of non-visiting customers in a plurality of evaluation areas included in each area category. Therefore, it is possible to more accurately evaluate the effectiveness of visit attraction in the evaluation target area when comparing areas with similar population sizes and similar characteristics as tourist destinations.

(7) Modification In addition, the system 1 may utilize trip data indicating the route from the start point to the end point of the user's behavior to perform more advanced analysis. As for behavioral analysis using trip data, it is possible to determine visits with higher accuracy by excluding terminal stations, which are included in trip data and serve as transportation nodes, from visit destinations.

Further, by determining the transportation means for each trip data, the tendency of visiting customers and non-visiting customers may be analyzed in detail for each access type (transportation means).

In system 1, as a determination of non-visit customers, among users who searched the evaluation area but did not visit, search with a search query such as "another area + sightseeing", and actually visit the area. It is also possible to judge only those who have visited the site as non-visit customers. Also, the other area may be specified as a competitive area.

Moreover, in the system 1, as for the attributes of non-visiting customers, for example, customer segments or preferences may be classified based on past search histories and behavior histories. For example, if the user previously searched for "stroller", it is possible to determine that the user belongs to the family segment and analyze that there is a demand for facilities for children.

Also, in the above embodiment, the action history and the search history are evaluated for the same evaluation period, but the present invention is not limited to this. For example, the system 1 separately sets a search evaluation period, which is a search history evaluation period, and an action evaluation period, which is an action history evaluation period, and evaluates the search history and the action history for different evaluation periods. may Also, the evaluation period may be set in advance, or may be set by the evaluator each time the evaluation area is specified.

Also, the area category may be classified according to either one of the evaluation index indicating the scale of the city and the tourist spot characteristics, or another index may be used to set the division of the area category. As another index, for example, accessibility from urban areas (time or cost required to move from urban areas to the evaluation area using general means of transportation, etc.) can be used.

In addition, the system 1 can be used not only for attracting tourists but also for various activities to attract customers. That is, by setting evaluation areas in units of shops and commercial facilities instead of the evaluation areas in units of cities described above, it is possible to use them for commercial activities of various scales. In other words, the system 1 explores the relationship between an evaluation object that shows interest in daily life and whether or not the user has taken an action toward the evaluation object, regardless of the content of the evaluation object.

For example, when the system 1 is used for purposes other than tourism attraction, an evaluation target category can be set according to the characteristics of the evaluation target instead of the tourist spot characteristics in the embodiment. Specifically, when evaluating various facilities, the following items can be adopted as evaluation target identification.
・Number of users and number of customers of the facility ・Purpose of use of the facility ・Accessibility of the facility ・Price range of the facility ・Classification of the facility (type of service provided)

Various keywords are assumed for the search target keyword, depending on what is to be evaluated. For example, when a search is performed with a search query such as "name of evaluation target +α", it can be determined as a customer candidate for the evaluation target. Here, various keywords corresponding to the content of the evaluation target can be considered as the search target keyword "α". If the evaluation target is a region, α = sightseeing, that is, "place name + sightseeing" can be used to determine whether or not the person is interested. can do.

Further, there is room for expanding the "name of the evaluation target" among the search queries such as "the name of the evaluation target + α" with respect to the determination of whether or not the customer is a candidate customer. For example, a customer who is interested in "Matsumoto City" may search using the search query "Matsumoto sightseeing" as well as the search query "Matsumoto castle sightseeing". It is also conceivable to use an abbreviation or common name of the name of the evaluation target in the search. For this reason, considering variations in notation for the name of the evaluation target, an assumed alternative expression may be set for each evaluation target in advance.

Further, in estimating the loss factor using the radar chart shown in FIG. 19, the classification division can be arbitrarily set for each area/facility/concept to be evaluated. For example, a sightseeing spot such as "Matsumoto" can be categorized as "Gourmet, Nature, History", but if a music event is to be evaluated, it can be categorized as "Access, Artist, Place". be done. In this way, an evaluation target category for classifying an evaluation target can be arbitrarily set according to the target.

Although the preferred embodiments of the present disclosure have been described above, the present disclosure is not limited to such specific embodiments, and the present disclosure includes the invention described in the claims and their equivalents. be Also, the configurations of the devices and the functions of the functional modules described in the above embodiments and modifications can be appropriately combined as long as there is no technical contradiction. Further, the order of the processes described in the above embodiment and modifications can be changed as appropriate as long as there is no technical contradiction.
Also, the program of the present invention may be expressed by multiple source codes, and the system 1 of the present invention may be realized by multiple hardware resources.

(9) Supplementary remarks The following are additional remarks regarding the embodiments of the present invention.

(Appendix 1)
A program comprising a computer having a processor and used in a system for analyzing customer behavior,
The program writes to the processor
receiving an input of an evaluation area to be evaluated;
a step of designating an evaluation area for a management server that manages a user's search history;
A mother set that is aggregated by referring to past search history and action history on the management server and contains information indicating the ratio of each area that actually visited during the evaluation period among the user groups that searched for the evaluation area during the evaluation period. obtaining data from a management server;
a step of determining a customer who has not visited the evaluation area during the evaluation period as a non-visiting customer among the customers included in the acquired population data;
a step of aggregating determined information related to non-visiting customers by referring to customer information managed in association with customer identification information.

(Appendix 2)
In the step of aggregating information related to non-visiting customers,
2. The program according to claim 1, which aggregates the percentage of non-visiting customers in the population data.

(Appendix 3)
In the step of aggregating information related to non-visiting customers,
2. The program according to claim 1, wherein customer segments included in non-visiting customers are classified based on any item managed as customer information, and customer distribution for each item is aggregated.

(Appendix 4)
In the step of aggregating information related to non-visiting customers,
2. The program according to claim 1, wherein search histories of non-visiting customers are referred to, and search queries with high search frequency among search queries including names of evaluation areas are totaled.

(Appendix 5)
In the step of aggregating information related to non-visiting customers,
2. The program according to claim 1, wherein the action history of the non-visiting customer is referred to, and the visited area instead of the evaluation area is identified as the competitive area.

(Appendix 6)
In the step of identifying a competitive area,
6. The program according to claim 5, wherein, among areas visited by non-visiting customers in place of the evaluation area, areas separated from the evaluation area by a predetermined distance or more are excluded from competition area candidates.

(Appendix 7)
to the processor, and
a step of determining a customer who visited the evaluation area during the evaluation period as a visiting customer, among the customers included in the acquired population data;
2. The program according to claim 1, causing execution of a step of referring to the behavior history of the visiting customer and identifying the area visited together with the visit to the evaluation area as the cooperative area.

(Appendix 8)
to the processor, and
a step of determining a customer who visited the evaluation area during the evaluation period as a visiting customer, among the customers included in the acquired population data;
2. The program according to claim 1, causing execution of the step of aggregating the difference in percentages of visiting customers and non-visiting customers.

(Appendix 9)
a step of referring to the search history of the non-visiting customer in the processor, and extracting the wording written together with the name of the evaluation area among the search queries as the search purpose keyword;
2. The program according to claim 1, causing execution of a step of aggregating information related to non-visit customers for each extracted search target keyword.

(Appendix 10)
In the step of extracting as a search target keyword,
9. The program according to claim 8, wherein, of the written words, words unrelated to tourism activities are removed as noise, and the remaining words are extracted as search target keywords.

(Appendix 11)
to the processor, and
a step of classifying the extracted search purpose keywords into keyword categories corresponding to tourism categories whose meanings correspond;
and aggregating information related to non-visiting customers for each keyword category.

(Appendix 12)
to the processor, and
10. A step of referring to the search history of visiting customers who have visited the evaluation area among the mother set data and outputting a keyword category corresponding to a search purpose keyword frequently appearing in a search query as strengths of the evaluation area. program described in .

(Appendix 13)
to the processor, and
a step of classifying each of the plurality of evaluation areas into a plurality of area categories set based on at least one of the evaluation index indicating the scale of the city and the characteristics of the tourist spot;
setting a reference value indicating the degree of non-visiting customers for each of a plurality of area categories;
2. The program according to claim 1, causing a step of performing evaluation based on a reference value for an evaluation area to be evaluated.

(Appendix 14)
13. The program according to claim 12, wherein the reference value is set by relatively evaluating the amount and ratio of non-visiting customers in a plurality of evaluation areas included in each area category.

(Appendix 15)
A method for use in a system for analyzing customer behavior comprising a computer having a processor, comprising:
The program is
receiving an input of an evaluation area to be evaluated;
a step of designating an evaluation area for a management server that manages a user's search history;
A mother set that is aggregated by referring to past search history and action history on the management server and contains information indicating the ratio of each area that actually visited during the evaluation period among the user groups that searched for the evaluation area during the evaluation period. obtaining data from a management server;
a step of determining a customer who has not visited the evaluation area during the evaluation period as a non-visiting customer among the customers included in the acquired population data;
and a step of aggregating the determined information related to the non-visiting customer with reference to customer information managed in association with the identification information of the customer.

(Appendix 16)
A system for analyzing customer behavior, comprising a computer having a processor, comprising:
The program is
means for receiving an input of an evaluation area to be evaluated;
means for designating an evaluation area for a management server that manages user search histories;
A mother set that is aggregated by referring to past search history and action history on the management server and contains information indicating the ratio of each area that actually visited during the evaluation period among the user groups that searched for the evaluation area during the evaluation period. means for obtaining data from a management server;
Means for determining a customer who has not visited the evaluation area during the evaluation period as a non-visiting customer among the customers included in the acquired population data;
and means for summarizing information related to determined non-visiting customers by referring to customer information managed in association with customer identification information.

1 customer behavior analysis system 10 customer terminal 20 evaluator terminal 30 behavior analysis server 301 communication unit 302 storage unit 3021 master set data 3022 non-visiting customer data 3023 visiting customer data 3024 total data 3025 reference value data 3026 evaluation result data 303 control unit 3030 Transmission/reception control module 3031 Judgment module 3032 Extraction module 3033 Aggregation module 3034 Classification module 3035 Reference value setting module 3036 Evaluation module 3037 Output module 40 Customer information management server 50 Customer database (DB)
60 action history database (DB)
70 Search history database (DB)
80 network 81 positioning device (GPS antenna)

Claims (15)

A program comprising a computer having a processor and used in a system for analyzing customer behavior,
The program causes the processor to:
receiving an input of an evaluation area to be evaluated;
a step of designating the evaluation area to a management server that manages a user's search history;
Includes information indicating the ratio of each area actually visited during the evaluation period, out of the group of users who searched for the evaluation area during the evaluation period, which is aggregated by referring to past search histories and action histories in the management server obtaining population data from the management server;
a step of determining a customer who has not visited the evaluation area during an evaluation period as a non-visiting customer among the customers included in the acquired population data;
a step of aggregating the determined information related to the non-visiting customer with reference to customer information managed in association with customer identification information ;
In the step of aggregating information related to the non-visiting customer,
A program for aggregating the proportion of the non-visiting customers in the population data . In the step of aggregating information related to the non-visiting customer,
2. The program according to claim 1, wherein customer segments included in said non-visiting customers are classified based on any item managed as said customer information, and customer distribution for each item is totaled. In the step of aggregating information related to the non-visiting customer,
2. The program according to claim 1, wherein said search history of said non-visiting customer is referred to, and said search queries with high search frequency among search queries including names of said evaluation areas are tabulated. In the step of aggregating information related to the non-visiting customer,
2. The program according to claim 1, wherein said action history of said non-visiting customer is referred to, and an area visited instead of said evaluation area is specified as a competition area. In the step of identifying the competitive area,
5. The program according to claim 4 , wherein, among areas visited by said non-visiting customers instead of said evaluation areas, areas separated by a predetermined distance or more from said evaluation areas are excluded from candidates for said competing areas. The processor further comprising:
a step of determining a customer who visited the evaluation area during an evaluation period as a visiting customer, among the customers included in the acquired population data;
2. The program according to claim 1, causing execution of a step of referring to the behavior history of the visiting customer and identifying the area visited together with the visit to the evaluation area as a cooperative area. The processor further comprising:
a step of determining a customer who visited the evaluation area during an evaluation period as a visiting customer, among the customers included in the acquired population data;
2. The program according to claim 1, causing execution of a step of aggregating a difference in percentages between said visiting customers and said non-visiting customers. a step in which the processor further refers to the search history of the non-visiting customer, and extracts, as a search target keyword, the words written together with the name of the evaluation area among the search queries;
2. The program according to claim 1, causing execution of a step of aggregating information related to said non-visiting customers for each of said extracted search purpose keywords. In the step of extracting as the search target keyword,
9. The program according to claim 8 , wherein out of said written words, words unrelated to tourism activities are removed as noise, and the remaining words are extracted as said search target keywords. The processor further comprising:
a step of classifying the extracted search target keywords into keyword categories corresponding to tourism categories corresponding to their meanings;
9. The program according to claim 8 , causing execution of the step of aggregating information related to said non-visiting customers for each of said keyword categories. The processor further comprising:
referring to the search history of visiting customers who have visited the evaluation area among the population data, and outputting the keyword category corresponding to the search purpose keyword frequently appearing in search queries as strengths of the evaluation area; 11. The program according to claim 10 , to be executed. The processor further comprising:
a step of classifying each of the plurality of evaluation areas into a plurality of area categories set based on at least one of the following: an evaluation index indicating the size of a city; and a characteristic of a tourist spot;
setting a reference value indicating the degree of non-visiting customers for each of the plurality of area categories;
2. The program according to claim 1, causing execution of a step of performing evaluation based on said reference value for said evaluation area to be evaluated. 13. The program according to claim 12 , wherein said reference value is set by relatively evaluating the amount and ratio of said non-visiting customers in said plurality of evaluation areas included in each of said area categories. A method for use in a system for analyzing customer behavior comprising a computer having a processor, comprising:
The method comprises: the processor;
receiving an input of an evaluation area to be evaluated;
a step of designating the evaluation area to a management server that manages a user's search history;
Includes information indicating the ratio of each area actually visited during the evaluation period, out of the group of users who searched for the evaluation area during the evaluation period, which is aggregated by referring to past search histories and action histories in the management server obtaining population data from the management server;
a step of determining a customer who has not visited the evaluation area during an evaluation period as a non-visiting customer among the customers included in the acquired population data;
a step of aggregating the determined information related to the non-visiting customer by referring to customer information managed in association with customer identification information ;
In the step of aggregating information related to the non-visiting customer,
A method of aggregating the proportion of the non-visiting customers in the population data . A system for analyzing customer behavior, comprising a computer having a processor, comprising:
The system, wherein the processor:
means for receiving an input of an evaluation area to be evaluated;
means for specifying the evaluation area for a management server that manages user search histories;
Includes information indicating the ratio of each area actually visited during the evaluation period, out of the group of users who searched for the evaluation area during the evaluation period, which is aggregated by referring to past search histories and action histories in the management server means for acquiring population data from the management server;
means for determining a customer who has not visited the evaluation area in an evaluation period as a non-visiting customer among the customers included in the acquired population data;
means for aggregating the determined information related to the non-visiting customer by referring to customer information managed in association with customer identification information ;
The means for aggregating information related to the non-visiting customer includes:
A system for aggregating the proportion of the non-visiting customers in the population data .

Images