Potential customer discovery support system

Abstract

We provide a potential customer discovery support system that has electronic information processing logic that calculates LSV values ​​based on lifestyles, calculates potential revenues for each town or block through profiling and rankings, and works in conjunction with map information to discover potential customers and support corporate sales activities. [Solution] The system comprises a database, an LSV navigation system that inputs lifestyle cluster data generated by a lifestyle cluster data generation unit based on statistical information, user-owned data of existing customers, and administrative boundary data, a calculation unit, and an input / output unit, calculates the LSV value of existing customers for each block, compares product contract performance for each block, creates a list of potential customers for areas with low product contract performance based on the LSV values ​​of areas with high product contract performance, and identifies potential customers by referring to the potential customer list.

Description

[Technical Field]

[0001] The present invention relates to a potential customer discovery support system with electronic information processing logic that efficiently supports the sales activities of financial institutions, companies, etc. by calculating LSV values ​​that indicate lifestyle-based value (LSV: Life Style Value) using information obtained by analyzing the trends of existing business customers, publicly available statistical information such as the national census, map information, and other various information related to sales, and by using the LSV values ​​to discover potential customers who are thought to belong to the same category or similar behavioral patterns. [Background technology]

[0002] There are two main patterns for sales activities by financial institutions, companies, etc. (hereafter simply referred to as "companies"). One is "additional sales to existing clients," and the other is "sales development to new customers" who are not existing customers. In the case of "additional sales to existing clients," salespeople can predict the customer's needs and propose products and services based on analysis of transaction history and customer information (age, gender, etc.). However, even for existing clients, there is a problem in that it is even more difficult to anticipate the needs of customers for whom the company is not the main customer (such as customers whose usage has stagnated), as there are no changes in transactions or new information is not accumulated.

[0003] Traditional marketing for individual product sales is based on the task of selecting potential customers. To do this, potential customer lists are extracted for each product as a separate task from the search function of a customer database (DB) provided in, for example, an MCIF (Marketing Customer Information File) or CRM (Customer Relationship Management).

[0004] Figure 1 shows the general procedure for conventional marketing. First, the customer database of MCIF or CRM is accessed (step S1), and target products are selected (step S2). Then, search criteria such as gender, age, average liquidity balance, and payroll (salary transfer) are used (step S3) to create a list of target customers (step S4). The listed target customers are accompanied by information such as the store's CIF (Customer Information File), account name, address, telephone number, and email address.

[0005] Next, a method for contacting the listed customers is selected (Step S5), which includes direct mail (DM), posting, inserts, e-mail, call centers (CC), ATMs (Automatic Teller Machines), and sales visits. After selecting the optimal method for contacting customers from these, the results of that contact method are confirmed (Step S6), and the marketing is completed by implementing the PDCA (plan-do-check-act) cycle (Step S7). For example, the hit rate (the percentage of direct mail orders that result in actual product purchases) that indicates the results after direct mail is generally said to be about 1% of the number of orders.

[0006] Because customer databases such as MCIF contain registered addresses, names, and other information, the output of this marketing method is the names of individual customers with whom the company does business. Therefore, various sales channels are used to search for these customer names. Whether the search terms properly identify target customers depends on the direct mail hit rate, which is typically low at around 1%. Furthermore, because the search terms themselves depend heavily on the know-how of the salesperson searching them, multiple salespeople often use similar search terms. Therefore, even if the PDCA cycle is repeatedly implemented after contact is made, there is very little room for improvement in marketing efficiency. Furthermore, because this marketing method is performed for each product, the same process is repeated for each product, resulting in inefficiency.

[0007] PDCA is widely known as a method of continuous improvement in business management such as quality control, and is a common method for continuously improving business operations by repeating the four steps of Plan → Do → Check → Action. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-096209 [Patent Document 2] Japanese Patent Application Laid-Open No. 2007-310851 [Patent Document 3] Japanese Patent Application Laid-Open No. 2004-118571 [Patent Document 4] Japanese Patent Application Publication No. 2020-194288 [Patent Document 5] Japanese Patent Publication No. 2020-149706 [Patent Document 6] Japanese Patent Application Publication No. 2019-179498 [Patent Document 7] Japanese Patent Application Laid-Open No. 2017-49784 [Patent Document 8] Patent No. 6729877 Summary of the Invention [Problem to be solved by the invention]

[0009] In conventional sales activities, LTV (Life Time Value) is used (steps S1 to S4 in FIG. 1), and the basic formula for calculating LTV is the following formula 1. (Number 1) LTV = Annual transaction amount x profitability x number of years of service (length of stay) However, this formula (1) is rarely used. This is because LTV differs for each customer, making it nearly impossible to calculate sales and expenses individually. In practice, it is often calculated based on the total number of customers (N). For example, the average LTV is calculated using formula (9) by calculating the number of repeat customers (formula (2)), customer revenue (formula (3)), customer retention costs (formula (4)), single-year profit (formula (5)), initial investment amount (formula (6)), total LTV (formula (7)), and LTV per customer (formula (8)). (Number 2) Number of repeat customers = Number of new customers x Customer retention rate (Number 3) Customer revenue = average customer price x number of repeat customers (Number 4) Customer retention cost = Number of new or repeat customers x Customer retention cost per person (Number 5) Single year profit = customer revenue - customer retention costs (Number 6) Initial investment = the cost spent to acquire customers in the first year (Number 7) Total LTV = Accumulated profit for a single year (Number 8) LTV per person = total LTV / number of new customers in the first year (Number 9) Average LTV = Additional LTV per person / N In order to maximize the LTV value, the key is to increase the "customer retention rate" in equation 2. If too much cost is spent, the "customer retention costs" shown in equation 4 will increase, which will reduce profitability.

[0010] Furthermore, LTV has many problems and accompanying conditions, as described below, and calculating LTV is an effective method if such problems can be resolved; however, if the problems, including the economic environment, cannot be resolved, calculating LSV (Life Style Value), which indicates value based on lifestyle, and profiling (analyzing a person's personal information and past behavior to predict future behavior, etc.) and ranking the LSV values ​​to calculate potential revenue (LSV navigation (Life Style Value navigation) (hereinafter simply referred to as "LSV navigation") is an effective way to support sales in uncovering potential customers. Here, a "potential customer" refers to a customer who does not yet feel the need for a product or service, or a customer who has not yet recognized a specific need. On the other hand, a customer who feels the need for a product or service is called a "prospective customer."

[0011] LTV refers to the benefits a customer brings to a company throughout their lifetime. Generally, companies with strong customer loyalty (i.e., loyalty) toward their products and services tend to have higher LTV values. Figure 2 shows a typical LTV along a time axis, illustrating the correlation between major events (leisure, marriage, home purchase, etc.) from entering the workforce to retirement and the financial needs associated with each event. Value (transaction amount) increases from entering the workforce to retirement until mid-life and then gradually decreases with age. The horizontal axis of events and financial needs represents the chronological flow of a lifetime. As shown in Equation 10, the sum of value and lifetime is LTV, which is the same as Equation 1. LSV values ​​will be discussed later. As shown in Figure 2, the LSV value of a customer at each time point is calculated for each event requiring funds (e.g., home purchase and mortgage). (Number 10) LTV = Lifetime + Value LTV values ​​calculated using mathematical formulas are merely pseudo-numbers. Accurately calculating the revenue from a single customer requires the accumulation of decades of transaction data, which is nearly impossible to obtain. Therefore, if we assume that people living in the same group share the same consumer behavior patterns, we can assume that the cumulative transaction results of people living in that group area with financial institutions will approximate the estimated LTV. To that extent, the LTV value calculated here represents cross-sectional revenue at a given point in time. Calculating revenue typically requires calculations of initial investment, asset value, sales and administrative expenses, etc. After performing these calculations, financial institutions maintain gross profit interest rates for each product to forecast revenue. Once these gross profit interest rates are provided, interest income on the remaining balance for each product can be calculated. However, the interest income calculated here is not the interest income of each individual, but the total of all accounts held in the group area. The other factor is the characteristics of individual products. For example, mortgage interest rates vary greatly depending on the loan application period, and since financial institutions have their own interest rate system, if they submit the weighted average interest rate of current mortgages, the proportion of high-interest mortgages will be of great significance. The higher the proportion of mortgages borrowed at low interest rates, as is the case now, the lower the gross interest rate on mortgages will be, and ultimately the lower the overall profits will be.

[0012] The present invention was made in light of the above-mentioned circumstances, and its purpose is to provide a potential customer discovery support system that has electronic information processing logic that calculates an LSV value that indicates value based on lifestyle, calculates potential revenue for each town or block through analysis such as profiling and ranking (LSV navigation), and discovers potential customers in conjunction with map information to support corporate sales activities.The purpose is to reuse existing customers as new customers for new products, to derive suitable customers as efficiently as possible, and to utilize this to contribute to reducing the costs of sales activities. [Means for solving the problem]

[0013] The present invention is a potential customer discovery support system that uses information processing logic to support the discovery of potential customers, and the above-mentioned object of the present invention is achieved by comprising a database, an LSV navigation system that inputs lifestyle cluster data generated by a lifestyle cluster data generation unit based on statistical information, user-owned data of existing customers, and administrative boundary data, a calculation unit, and an input / output unit, calculating the LSV value of the existing customers for each block, comparing product contract performance for each block, creating a potential customer list for areas with low product contract performance based on the LSV values ​​of areas with high product contract performance, and discovering potential customers by referring to the potential customer list. [Effects of the Invention]

[0014] In this invention, statistical information such as demographic data published in censuses and the like is used to group (segment) by town or chome, and calculates an LSV value that indicates the value based on each person's lifestyle (LSV: Life Style Value) for each town or chome, and also calculates potential revenue including the number of potential customers using the total number of accounts or total population of the town or chome. If LSV values ​​were calculated on a city, town, or village basis, the target area would be too broad and it would be impossible to calculate an LSV value with a high probability, so in this invention, LSV values ​​are calculated on a town or chome basis.

[0015] The potential customer discovery support system using LSV Navigator of this invention is a method to find customers who can do business with you over the long term with the new products you want to offer with low risk.To achieve this, this invention aims to scientifically analyze and evaluate the components (elements) of "luck, ability, and connections" that customers possess on the past, present, and future time axes.

[0016] If we imagine a set represented by a finite group, the components of "luck, ability, and connections" include lineage, land, assets, family, relatives, and friends, the Feng Shui of the land, one's mother's genes and luck, abilities, and connections, and one's father's genes and luck, abilities, and connections. The neighborhood is also one of the components, and the "luck, ability, and connections" of each individual are developed by the limited components such as where one lives and family background, along with the multiple components that create the people who live in that neighborhood. As a result, the idea is that a person's "luck, ability, and connections" change over time due to the components within them as they grow.

[0017] The components of "luck, ability, and personal connections" do not have quantities or units, and many are expressed in natural language, raising the question of whether each item in Figures 16, 17, and 18 is a discrete or continuous quantity. The difference between discrete and continuous quantities is that discrete quantities are counted, while continuous quantities are measured. In order to resolve the issue of determining the unit value of each item used to evaluate an individual, the items for "luck" are "savings, real estate, and occupation," "ability" are "credit history and income," and "personal connections" are "family structure and age," and the risk items of "interest, purchasing power, and age" are used as dependent variables, and analysis is performed by adding items related to "interest" to luck, "purchasing power" to ability, and "age" to personal connections.

[0018] In order to improve on traditional inefficient marketing methods, publicly available statistical information such as the national census is used to calculate LSV values ​​based on lifestyle cluster data (geographical demographic data) that shows consumer behavior characteristics by area generated by clustering analysis, large amounts of customer information (CRM / SFA (Sales Force Automation)) data held by companies, administrative boundary data including population and area data, and the lifestyles of existing customers.The LSV values ​​are calculated by reflecting risk values ​​obtained by analyzing the three major risk factors of existing customers: purchasing power, interests, and age.

[0019] By calculating an LSV value that takes into account risk values ​​based on the three major risk factors of purchasing power, interests, and age, potential customers with similar rankings and high LSV values ​​are pinpointed and selected in order from the top of the ranking. For the selected potential customers, potential profits (including the number of potential customers) for each area of ​​personal products are calculated, promoting and supporting sales of personal products with a higher probability than before. This makes it possible to provide an efficient potential customer discovery support system based on the customer's LSV value to financial institutions and companies seeking to discover potential customers in advance.

[0020] The LSV value is a numerical value that indicates the similarity of lifestyles, and the LSV Navigator of the present invention is a tool that can be used in potential customer discovery support systems that perform profiling and rankings using LSV values ​​derived from a predetermined calculation formula, in GIS (geographic information system) flow businesses (businesses that make a profit by selling out with each transaction), and in stock businesses (businesses that make continuous profits by entering into contracts with customers and securing members) that provide customer rankings and calculations of the value (potential profits) of towns or districts.

[0021] Furthermore, by displaying the discovered potential customer information on a residential map or a wide-area map, it is possible to obtain regional characteristics for each branch of a financial institution or company, etc. Regional characteristics include population size, population density, area division, climate, and transportation conditions, making it an effective tool for visualization.

[0022] The potential customer discovery support system of the present invention expresses different multifaceted elements as diversity and maps them onto an intuitively understandable Euclidean plane according to the purpose of use, and can be widely used not only by financial institutions such as banks but also by companies in other industries such as sales companies.In addition, using statistical information such as the national census conducted every five years, consumer behavior characteristics by area generated by clustering analysis, and target customer information obtained from customer transaction data accumulated in CRM / SFA owned by companies can be shared and used with other systems. [Brief explanation of the drawings]

[0023] [Figure 1] 1 is a flowchart illustrating a conventional marketing procedure. [Figure 2] FIG. 1 is a schematic diagram for explaining LTV. [Figure 3] 1 is a schematic flowchart showing the principle of LSV navigation used in the present invention. [Figure 4] FIG. 10 is a diagram showing a flow of target extraction by LSV navigation. [Figure 5] FIG. 10 is a diagram showing an example of an analysis result report. [Figure 6] 1 is a block diagram showing a configuration example of the present invention; [Figure 7] 1 is a flowchart showing an example of the overall operation of the present invention. [Figure 8] 10 is a flowchart illustrating an example of calculating the total product amount of potential customers. [Figure 9] 10 is a flowchart illustrating an example of the operation of tree analysis. [Figure 10] FIG. 1 is a diagram illustrating an example of a tree. [Figure 11] 10 is a flowchart illustrating an example of an operation for calculating the probability of a product contract. [Figure 12] 10 is a flowchart illustrating an example of the operation of a potential customer extraction process. [Figure 13] 1 is a diagram showing an example of the relationship between living standards and lifestyle groups. [Figure 14] 1 is a diagram showing an example of lifestyle codes and segments. [Figure 15] This is a map showing an example of the relationship between addresses and living standards. [Figure 16] FIG. 10 is a schematic diagram of risk value calculation regarding purchasing power. [Figure 17] FIG. 10 is a schematic diagram of interest-related risk value calculation. [Figure 18] FIG. 1 is a schematic diagram of calculation of risk values ​​related to age. DETAILED DESCRIPTION OF THE INVENTION

[0024] This system analyzes existing customers' lifestyles to calculate similarities in lifestyle cluster data, analyzes risk values ​​related to risk factors such as age, and calculates numerical LSV values. It then references the LSV values ​​for products designated as sales targets to efficiently pinpoint potential customers in other regions. Specifically, this system expresses diverse, multifaceted elements as diversity and maps them onto an intuitively understandable Euclidean plane tailored to the intended use. It analyzes lifestyles using existing customer data (user-owned data) stored in a company's CRM, SFA, MCIF, etc., publicly available statistical information such as the census, and publicly available administrative boundary data, including topographical data, to calculate lifestyle cluster data and determine similarities between the lifestyle cluster data. It segments store VCIF accounts by town or block (hereinafter simply referred to as "town / block") and calculates potential revenue and risk values ​​for products designated (input) as sales targets. Then, based on the similarity of lifestyle cluster data, potential profit and risk value, an LSV value indicating value based on lifestyle for each block is calculated, customers with a high record of product contracts on a block-by-block basis are identified, the potential customer rate for another block-by-block basis is calculated, and the potential profit for each potential customer is estimated using the total number of accounts or total population from the total number of products for the potential customer. This is an innovative sales support system that can strongly promote sales of products for individuals at companies such as banks and scientifically contribute to the discovery of potential customers in other regions where those products are being purchased.

[0025] Here, "potential revenue" refers to the revenue predicted when potential customers and prospective customers are added together and future customers realize they need a product or service and become customers, and is calculated using the following equation (11). The "number of potential customers" is the predicted number of customers who do not yet realize they need a product or service, or who realize they need it but have not yet introduced or purchased it (the predicted number including potential customers and prospective customers). The number of potential customers is calculated using the following equation (12). (Number 11) Potential revenue = (annual income or annual transaction amount of the town / district using statistical information) / (total population or total number of accounts of the town / district) (Number 12) Number of potential customers = total number of households or total population of the town / district - number of existing customer accounts or number of existing customers after name aggregation The "contract rate for potential customers" is calculated by calculating the highest contract rate for existing customers in the block according to the following formula 13, subtracting the contract rate for existing customers in each block, and dividing the result by the highest contract rate for existing customers in the block. In other words, the contract rate for potential customers is calculated according to the following formula 14. (Number 13) Contract rate for existing customers = (Number of existing customer accounts or customers after name matching by town / district) / (Total number of households or total population of the town / district) (Number 14) Contract rate for potential customers = (Contract rate for the highest existing customers in the block - Contract rate for existing customers in each block) / (Contract rate for the highest existing customers in the block) In this invention, potential profits are calculated for each block. Risks due to technological advances, which significantly affect values, and the influence of the economy, politics, and the global economy are objectively quantified as risk values ​​based on risk factors (e.g., purchasing power, interests, age) based on personal characteristics. The risk values ​​are then reviewed annually (e.g., every 1 to 5 years). In this invention, risk values ​​are calculated based on purchasing power, which is likely to be linked to luck; interests, which are likely to be linked to ability; and age, which is likely to be linked to personal connections. The creditworthiness of the invisible individuals of existing customers (e.g., status characteristics, ability to perform, credit history, personal connections, behavioral preferences, etc.) is inversely proportional to the risk value; the higher the risk value, the lower the creditworthiness; and conversely, the lower the risk value, the higher the creditworthiness. This relationship is shown in Equation 15, described below. In this invention, risk values, which are easy to calculate numerically, are calculated, and the LSV value is calculated using the risk values.

[0026] In current markets, such as those related to the Sustainable Development Goals (SDGs), the automobile industry, and the shipbuilding industry, many approaches are being adopted to reinvent themselves by creating new rules. In other words, rather than seeking new markets (industries), these approaches are often based on reusing existing markets. To more effectively propose new products to existing markets, a system is needed that calculates customer value closer to the current situation based on existing customer lifetime value (LTV). However, as mentioned above, there are problems with assessing LTV. Therefore, this invention proposes a new LSV navigation system that uses LSV values ​​to analyze the lifestyles of existing customers, calculate their lifestyle value (LSV value), and discover potential customers in areas with poor sales performance. To give an example of the LSV navigation of the present invention, it is a system (navigation) that uses publicly available statistical information such as the national census and administrative boundary data (topographical data), and utilizes lifestyle cluster data for each area generated by clustering classification as an example of information on good fishing grounds, while also extracting and acquiring (pinpointing and "hitting") target customers (potential customers) from customer data accumulated in CRM / SFA etc. owned by financial institutions and other companies.

[0027] FIG. 3 is a schematic flowchart showing an outline of the LSV navigation used in the present invention. As shown in FIG. 3(A), first, targets (customers who will purchase the product being the sales target) are determined, and then, as shown in FIG. 2(B), the block / town where the target lives (an analogy of a good fishing spot) is identified. For example, it may be an area where many wealthy people live, which is common to areas with high purchasing volumes of the product being the sales target. Then, using the LSV navigation, the target is narrowed down as shown in FIG. 3(C), and then, as shown in FIG. 3(D), a list is collated based on a database (DB) analysis, and targets (potential customers) are acquired in the sales area, that is, one-on-one fishing (hits) are conducted to promote sales of the product (FIG. 3(E)).

[0028] Figure 4 shows the process for discovering targets (potential customers) using LSV Navigation. User data held by user companies such as financial institutions is classified with information such as customer number, name, gender, address, and date of birth, as shown in Figure 4(A). This user data is combined with lifestyle cluster data LD obtained by lifestyle cluster classification, as shown in Figure 4(B), and administrative boundary data, as shown in Figure 4(C), to calculate LSV values, as shown in Figure 4(D), according to equation (16). The LSV values ​​are then used to generate an analyzed potential customer list (ranking), as shown in Figure 4(E), for the area where sales are planned. The LSV value is calculated by calculating the similarity of the lifestyle cluster data LD, calculating potential revenue from the number of potential customers in each block, and multiplying the similarity of the lifestyle cluster data LD, potential revenue, and risk value for each customer. "Similarity" in lifestyle cluster data LD means that people living in neighborhoods within groups classified by consumer behavior characteristics (geographical demographic data) using publicly available data information such as the national census are expected to exhibit essentially the same consumer behavior, as shown in the lifestyles in Figures 13 and 14. As there are no regional differences in ordinary deposits and fixed-term deposits, which are representative of assets held in custody, it is thought that there are similarities in consumer behavior patterns on a neighborhood-by-chome basis.

[0029] However, if the risk values ​​are multiplied directly, the larger the risk value, the larger the LSV value, so they are multiplied by a value converted to creditworthiness. In other words, since risk and credit are inversely related, if the risk value and creditworthiness are expressed as a maximum value of "100(%)", the risk value is expressed by the following formula 15. (Number 15) Risk value (%) = 100 - Creditworthiness → Creditworthiness = 100 - Risk value (%) Therefore, the LSV value is calculated using the following equation 16. (Number 16) LSV value = lifestyle cluster data LD similarity x potential revenue ×(100-risk value (%)) Figure 5 shows details of an individual customer (e.g., Rank 1) in the potential customer list, which shows the analysis results. The potential customer list includes information such as rank, name, group, segment, and risk value. The risk value is a numerical value representing risk factors based on a person's characteristics, such as purchasing power, interests, and age, and can be thought of as creditworthiness. The lifestyle cluster data LD, consisting of pairs of groups (codes) and segments, is assigned a lifestyle code (described below). Based on the assumption that people living nearby have similar attributes and status, and share similar lifestyles, behaviors, and attitudes, the lifestyle cluster data generation unit (210) performs clustering based on statistical information such as the census. As will be described in more detail later, clustering can be performed, for example, with Group A consisting of elites active in large cities (highly educated, high-income elites who live in high-rise apartments in urban areas and drive Japan's economy and culture), and Group B consisting of executives in high-end residential areas (wealthy families who live in high-end residential areas on the outskirts of large cities and have achieved social status through promotion at large companies). In addition, the administrative boundary data includes population and area for each block.

[0030] Figure 6 shows an example of the overall configuration of the system 100 of the present invention in cooperation with the system. The potential customer discovery support system 100 of the present invention comprises a DB (database) 110 and an LSV navigation system 120 that cooperates with the DB 110, and like a general computer, has an input unit 111 such as a keyboard and mouse, an output unit 112 such as data output, transmission, printer, etc., a display unit 113 such as a display screen, and a calculation unit 114 consisting of a CPU, memory, etc. The LSV navigation system 120 is composed of an LSV value calculation unit 121 that calculates LSV values, a ranking unit 122 that ranks potential customers, a profiling unit 123 that assigns lifestyle raster data to potential customers and makes estimates, and a segmentation unit 124 that groups them into towns and districts.

[0031] The LSV navigation system 120 receives customer data from the user-owned data 200, lifestyle cluster data LD from the lifestyle cluster data generator 210, and administrative boundary data 220. The administrative boundary data 220 is GIS data for administrative boundaries across the country, including prefecture names, branch offices and regional bureau names, counties, designated cities, municipalities, and administrative district codes. The lifestyle cluster data LD is generated by the lifestyle cluster data generator 210 from statistical information such as the national census. Generally, "statistical information" is the aggregated and processed results of surveys conducted under certain conditions, such as the national census and commercial statistics. However, the statistical information 250 used in this invention includes information surveyed by various government ministries and agencies (including the national census), local governments, and various industry associations. For example, vegetables = statistical information, cabbage = national census, and radish = commercial statistics. The divisions commonly used in existing area marketing systems are "mesh" and "administrative boundaries." In this invention, if we apply this to Figure 3, the area showing good fishing grounds is the "administrative boundary," and the statistics that examine what kind of fish are present within that area are the "census."

[0032] A customer 200, such as a financial institution, is equipped with a CRM / SFA 205 and has user-owned data including individual customer data 201, corporate customer data 202, transaction data 203, and product data 204. Sales Force Automation (SFA) is a sales support system, and CRM includes support and management of sales activities. Therefore, CRM / SFA 205 includes SFA as one of the components that implements CRM. More precisely, CRM, which is used for a wide range of tasks from marketing to support services, includes SFA, which supports sales, as a function. User-owned data is held by user companies such as financial institutions. In the case of financial institutions, the data is stored in host computers and various systems (including CRM and SFA) according to acquisition purpose, and can be extracted as needed as individual customer data 201, corporate customer data 202, transaction data 203, product data 204, and other data. These user-owned data groups are imported into the LSV navigation system 120, and the calculated LSV values ​​are returned to the CRM / SFA 205. However, if the user already owns the CRM / SFA 205, as in this example, the values ​​are stored in the temporary storage DB 110.

[0033] The LSV navigation system 120 calculates an LSV value based on equation (16) for various data (individual customer data 201, corporate customer data 202, transaction data 203, product data 204) held by a user (purchaser or user) using lifestyle cluster data LD from the lifestyle cluster data generation unit 210 and administrative boundary data 220, and creates a ranked potential customer list by a ranking unit 122, which is output from the output unit 112 and displayed on the display unit 113. The LSV value calculation unit 121 is a collection of statistical methods and tools that support marketing activities, and calculates an LSV value by analyzing various acquired data. The segmentation unit 124 subdivides the market (by block) and groups customer segments with similar characteristics (ways of thinking and values) divided from a certain market by needs (products to be sold (e.g., new products, products the customer does not own, etc.)). Furthermore, the ranking unit 122 ranks the obtained collection based on the LSV values, and the profiling unit 123 estimates lifestyle cluster data for potential customers who could become customers of the company's products or services.

[0034] Figure 7 shows a sales support flowchart for a scenario in which Regional Bank A introduces "demographic data by block." The LSV navigation system 120 first inputs lifestyle cluster data LD, which is a group of blocks, and then imports customer data for blocks belonging to Regional Bank A from user-owned data (step S10). Next, the segmentation unit 124 groups (segments) the store CIF (store-customer information file) by block (step S11). After name matching, the age of the store CIF account holder and the lifestyle cluster data LD for each group are aggregated (step S12). A CIF is a unique management number assigned to each customer when they open an account. This number is linked to a branch number that identifies the financial institution's branch (branch) to create a store CIF. "Name matching" refers to the process of integrating data from multiple distributed databases into a single entity, such as by assigning the same ID to the same person, company, or household.

[0035] Next, the LSV value calculation unit 121 calculates the LSV value for each customer using Equation 16 (step S20). Furthermore, the target product (the product to be sold (e.g., "5-year fixed term deposit of 1 million yen")) is set via the input unit 111 for each age group based on the risk value used to calculate the LSV value (step S30). As in Equation 16, the LSV value is calculated by finding the similarity of the lifestyle cluster data LD, calculating potential profits from the number of potential customers for each block, and multiplying this by the risk value for each customer. However, since there are risk factors due to the characteristics of each individual (existing customer), risk values ​​that quantify customer risk for, for example, purchasing power, interests, and age are used. The LSV value is calculated for all customers in each block, using profit calculations (based on net business profits), potential profits, and risk values ​​for each age group, group, and product. The calculation of the LSV value and risk value will be described later.

[0036] Then, product contract records for each block (e.g., area X) are extracted for the set product from the user-owned data based on the calculated LSV values ​​(step S31), and the ranking unit 122 determines and ranks customers with high product contract records for each block (step S32). Since the level of product contract records for each block is known and the LSV values ​​of all customers are also known, the LSV values ​​of customers with high product contract records can also be determined. The calculation unit 114 determines the expected contract rate for potential customers based on the difference between the contract rates of existing customers in the area with the highest product contract record and the area with the lowest product contract record (step S33), thereby determining the expected number of potential customers for each block. The expected contract rate for potential customers is determined according to the following equation 17. (Number 17) Contract likelihood rate for potential customers = Contract rate for the highest existing customers in the town / district - Contract rate for existing customers in areas with low contract performance After the contract likelihood rate of potential customers is calculated by Equation 17, the maximum value of the number of potential customers is calculated by multiplying the number of potential customers excluding existing customers in the block by the contract likelihood rate of potential customers. (Number 18) Number of potential customers = Number of potential customers x Contract probability rate of potential customers Next, the total product total of the number of prospective potential customers is calculated by aggregating all groups in the same age category, all age categories, and all products (step S40), and the profiling unit 123 calculates the potential profit of the number of prospective potential customers according to equation 11 (step S50), and sales (normal sales operations) are carried out based on the number of prospective potential customers by block (step S51). By calculating the potential profit and understanding the difference with the profit of the actual business partner, market value can be determined, such as whether to uncover existing transactions or reconsider the company as a source of new products. The estimation of the potential profit of the number of prospective potential customers in equation 11 is carried out by comparing it with the current number of product accounts to find profit using the LSV value.

[0037] Figure 8 shows the details of step S40. First, the combination of attribute and transaction items for the account with the highest contract rate is examined, and the contract rate by product and group is predicted (step S41). Then, the population is identified based on the accounts in the corresponding block that have the product and the accounts in the block that do not have the product (step S42). That is, to identify the population, the product, age group, gender, and group are first identified. Once the group is identified, the store CIF accounts (hereafter simply referred to as "accounts") that have the product after name aggregation within the block are tallied, along with all accounts in the same block that do not have the product. The total number of both accounts forms the population, and if this is less than 5,000, the identification criteria are relaxed. For example, if there are not 5,000 accounts in the age group (in 5-year increments), the age group is narrowed to 10-year increments, and if necessary, certain criteria are completely removed. The attribute information and transaction items of the accounts are determined in advance. For example, when assigning a group to a store CIF (after name matching), a group is assigned to the store CIF after name matching, and the account is assigned the mosaic group name, name, age, gender, address, and transaction items, and the account attribute information items and transaction items are determined in advance and are the same for all accounts.

[0038] Next, a tree analysis is performed (step S43), and the results of the tree analysis are applied to each block / district account (step S44). Details of the tree analysis will be described later. To apply the results to each block / district account, a ranking list is created for accounts that have the combination of transaction items obtained from the tree analysis for each block / district and do not currently own the product, sorted in descending order of contract rate, and a list of potential customers with a high probability of subscribing to the product and the number of customers is obtained. The potential customer list (see Figure 4(E)) includes information such as the store CIF, name, address, and telephone number. After applying the results to each block / district account in step S44, an individual rating is assigned to the potential customer list (step S45). The individual rating serves as a database of individual customer risk estimates, enabling simultaneous sales promotion and risk control. In this case, even if a potential customer has a high probability of subscribing to a product, customers with a high risk value (e.g., a risk value of 95 or higher) must be excluded before contacting them. Similarly, accounts with prohibited contacts should also be excluded in advance. For loan products, accounts with a high contract rate and low personal ratings have the highest contract rate. For assets held in custody products, accounts with a high contract rate and high personal ratings have the highest contract rate. In either case, accounts with a high risk value are excluded in advance.

[0039] Details of the tree analysis in step S43 above will be explained with reference to the tree diagram in Figure 10. The population (total number of samples: "1000", total number of accounts: "1500") is first classified into males (total number of samples: "6000", total number of accounts: "1200") and females (total number of samples: "4000", total number of accounts: "300"), and both males and females are further classified by whether or not they have a credit card, with nodes "001" to "004" assigned to each. The parentheses for each item in Figure 10 indicate the number of classified samples, the number of accounts with a credit card, and the percentage of accounts with a credit card. All variables are AND-conditions, so node "001" represents accounts for males with a bank credit card, node "002" represents accounts for males without a bank credit card, node "003" represents accounts for females with a bank credit card, and node "004" represents accounts for females without a bank credit card. In reality, variables (total number of samples and total number of accounts) appear one after another, but here we use two variables, and combinations of variables are called nodes, with four nodes shown here ("001" to "004"). Each node displays the combination of variables and the contract rate (card ownership rate) for credit card products indicated by that combination. Almost all financial institutions issue customers with cash cards that also serve as credit cards, but because there are also customers who do not use credit, the card ownership rate is not zero even if they do not have a credit card.

[0040] In the tree analysis shown in Figure 10, the highest credit card contract rate is for accounts at node "001" where a male has a credit card, with a high credit card contract rate of 27.5%. On the other hand, for accounts at node "004" where a female does not have a credit card, the credit card contract rate is a very low 1.1%. After the tree analysis is complete, the nodes are sorted in descending order of contract rate, and the combination of attribute items and transaction items is determined for each node. In the example of Figure 10, the order of highest contract rate is (1) male with a credit card, (2) female with a credit card, (3) male without a credit card, and (4) female without a credit card.

[0041] The details of step S44 are shown in the flowchart in Figure 9, which describes the process of applying the tree analysis results to each block / district account. First, the probability of signing up for a product is calculated (step S44-1), and risk (or credit) is analyzed based on the combination of attribute and transaction items (step S44-2). The method for calculating the probability of signing up for a product is described later, but it uses lifestyle codes and user-owned data. Among the accounts in each block / district in the same group, accounts with the same attribute and transaction items as the node with the highest contract rate, but which do not own a credit card product, are selected and listed. The selection is then performed sequentially in descending order of the node's contract rate (step S44-3), and the combination of transaction information for each node is determined (step S44-4). The selected accounts are specific accounts with names, addresses, and phone numbers. Therefore, whether the contact is a direct mailer or a sales representative, the sales activity is directed at a specific individual. Here, a personal rating is further assigned to the customer list of the listed accounts. When assigning the personal rating, customers with a risk value of 95 or higher are excluded from the list.

[0042] An example of the operation of the method for calculating the probability of a product contract will now be described with reference to the flowchart of FIG.

[0043] First, the accounts are classified into groups from the account list based on the account's street address, age, contract rate for each product in the street address, etc. (Step 44A), and the product is identified (Step 44B). Next, the contract probability 1 for a specific product determined among accounts in the same group is calculated (Step 44C), and the contract probability 2 for a specific product is calculated by treating accounts with the specific product and accounts without the specific product as one population (Step 44D). After that, tree analysis is used (Step 44E), and the contract probability for a specific product is displayed in the terminal node (Step 44F). The terminal node also displays the number of samples and the number of contracted accounts contained in that node, and the ratio is output as the contract probability (Step 44G).

[0044] The potential customer extraction process performed by the potential customer discovery support system 100 will be described in detail with reference to the flowchart of FIG. 12 (a detailed version of FIG. 7).

[0045] When a customer holds multiple accounts, customer attribute information (such as name, address, date of birth (individual), date of establishment (corporate), etc.) is used to verify attribute matches with existing accounts. If there are matches, the accounts must be "centrally managed" as multiple accounts belonging to the same customer. After the name matching process for this centralized management (step S100), the segmentation unit 124 assigns groups according to the block / town / district of the account holder's address (step S110) and categorizes the accounts by age, for example, in five-year increments (step S111). The accounts categorized by age are further grouped based on the consumer behavior similarity (lifestyle cluster data LD) generated by clustering analysis, which uses external data for purchasing, and the area-based classification results obtained from lifestyle codes and user-owned data (step S112). The accounts in the groups are then again sorted by block / town / district (step S113). By combining the external purchasing data with customer transaction data stored in the CRM / SFA 110, LSV values ​​(albeit pseudo-values) based on major individual products are generated. The generated LSV values ​​are by definition required to be maximized, and this maximization process is carried out. The key purpose of the LSV navigation system is to increase "potential customer discovery" and "customer retention rate." Spending too much money increases "customer retention costs," which reduces profitability. It is important to increase "potential customer discovery" and "customer retention rate (contract utilization rate)" while keeping costs down by streamlining operations.

[0046] In order to increase the transaction price for each town / district for targets residing in the identified sales area, potential customers are predicted by maximizing purchase frequency and the possibility of converting existing customers into new customers. In other words, differences with current customers are confirmed. Based on the LSV values ​​obtained in this process, the size of target customers for each area for all personal products is calculated. Then, the total number of accounts and the number of product contracts for each town / district is calculated (step S114), the population of each town / district is calculated (step S115), and the contract rate for each product for each town / district is calculated (step S116). This contract rate is the contract rate per capita or per account.

[0047] Next, for each product, the contract rate of the block with the highest contract rate is determined (step S117), and the contract rate of each block is subtracted from the highest contract rate to calculate the expected contract rate of potential customers (step S118).The expected contract rate of potential customers is then multiplied by the population or number of accounts in that block to determine the number of expected potential customers (step S119).Even if the number of expected potential customers is determined, normal sales efforts are required to turn them into actual customers.The above steps S110 to S119 are the classification process.

[0048] From the account list obtained in step S110 above, potential customers for individual products are classified into groups for each area (step S120), and the products are identified (step S121). Next, the contract rate for the specific product determined among accounts in the same group is calculated (step S122), and the contract rate for the specific product is calculated for accounts that have the specific product and accounts that do not have the specific product as a single population (step S123), and a tree analysis as shown in Fig. 10 is performed (step S124). The contract rate for the specific product is displayed as a terminal node, and the node indicates the number of samples and the number of contracted accounts included in the node, as shown in Fig. 10, and the ratio is displayed as the contract rate.

[0049] Next, all nodes are sorted in descending order of contract rate (step S125), and the number of accounts without a specific product among the sample of nodes with the highest contract rate is displayed (step S126). Next, accounts without a specific product are classified and listed by block (step S127), and the above method is applied sequentially to each node with a high contract rate (step S130). Accounts within each block group are listed in descending order of the contract rate for specific approval (step S131), and the above method is applied to all groups (step S132), and then to all products (step S133). With this method, all accounts without a specific product may potentially be subject to contract. However, it is important to specify accounts with a high contract rate within the current transaction situation. Therefore, the accounts listed in descending order of contract rate in each block must be multiplied by a cap, and this cap is the number of potential customers for each block shown in step S127. The number of potential customers is calculated for each age group. The accounts listed by block are reclassified by age group (step S134), and the results are compared with the number of potential customers. The final output is a list of accounts with a high contract rate for purchasing the product, by group, product, and block (step S140).

[0050] A known classification of lifestyles, including living standards, is shown in Figure 13, where urbanization levels are categorized into "rural area," "suburban area," "city area," and "inner city area," and wealth levels are also defined for each of these. For example, wealth levels for "inner city area" are categorized into "trendy lifestyle," "urban middle class," and "downtown." A list of lifestyle cluster data assigned to each classification is shown in Figure 13, and the relationship between a customer's address and their living standards is also determined by area information, such as that shown in Figure 15. Therefore, lifestyle codes can be automatically assigned based on this information.

[0051] The contents of the lifestyle cluster data LD are shown in FIG. 14, for example, and the lifestyle cluster data is composed of a lifestyle code and a segment.

[0052] Next, the analysis items that are taken up as data that companies are likely to possess include, for example, savings and investments, occupation, real estate, credit history, income, family structure, and age, and each item is associated with a vector space, which also corresponds to map information indicating companies and personal names (related to family tree information), including individuals.

[0053] Figure 16 shows an example of risk calculation for purchasing power, one of the three major risks. The characteristic factors are educational background, occupation, income, credit, family structure, hobbies, and real estate. Educational background includes whether or not a person studied abroad and their highest level of education (vocational school, high school, university, etc.). Occupation includes manager (founder, successor, etc.), size, and employee (large company, civil servant, small or medium-sized enterprise, etc.). Income includes age, dividend income, and salary. Family structure includes age groups, such as two or more households living together, nuclear family, and single. Hobbies include luxury cars and yachts, travel, relationships with women (rumored, not rumored, unknown, etc.), and gambling (good luck, likes, average, doesn't do, etc.). Real estate includes inheritance (business land, home, farmland, rental apartment, etc.), self-acquisition (business land, home, farmland, rental apartment, etc.), and real estate rental (business land, farmland, rental apartment, etc.). Credit history includes good, fair, poor, and anti-social affiliation.

[0054] The products that are subject to risk analysis are asset management (investment trusts), life insurance, liquid deposits (foreign currency deposits, fixed term deposits, etc.), mortgages (individuals, corporations), and credit cards, and it is also applied to the sale of goods. An example of risk calculation for the risk factor interest is shown in Figure 17, and an example of risk calculation for age is shown in Figure 18. Investment in Figure 17 includes investment trusts, deposits, and stocks with a track record, interest, and none, while health in Figure 18 includes illness / disability due to aging, chronic illness, disability, and health.

[0055] If "interest = no interest" and the assumed risk value is "90," we calculate the value converted into creditworthiness. If it's better to give up or you're unsure, the risk value is assumed to be 40, and if you're interested, the risk value is assumed to be 15, which is converted into creditworthiness. Also, "purchasing power = no money," which is a factor that includes education, occupation, income, credit, family structure, hobbies, and real estate. "Age = no chance," where age is the risk value, and the older you are, the more difficult it is to recover and the more chance there is. The three major risks are (A) no interest, (B) no money to buy, and (C) being too old to take chances, and the worst risk occurs when all three factors are present. Creditworthiness is calculated using combinations of the above three factors (A, B, C, AB, AC, BC, ABC). For example, "no money, but interested and young," "young and money, but not interested," or "money and interested, but no chance." [Explanation of symbols]

[0056] 100 Potential Customer Discovery Support System 110 DB (Database) 111 Input section 112 Output section 113 Display section 114 Arithmetic section 120 LSV navigation system 121 LSV value calculation unit 122 Ranking Section 123 Profiling Department 124 Segmentation Department 200 Customer System 201 Personal Customer Data 202 Corporate Customer Data 203 Transaction Data 204 Product Data 205 CRM / SFA 210 Lifestyle Cluster Data Generation Department 220 Administrative Boundary Data 250 Statistics

Claims

1. A method for supporting potential customer discovery using a support system in which a database is input; lifestyle cluster data for each town or block, which is generated by clustering classification in a lifestyle cluster data generation unit based on statistical information compiled and processed from the results of surveys conducted under certain conditions by various ministries and agencies, local governments, and various industry associations, etc., and which is publicly available; user-owned data including personal customer data, corporate customer data, transaction data, and product data; and publicly available administrative boundary data for the entire country, which is compiled as GIS (geographic information system) data, including prefecture names, branch office / regional bureau names, counties, designated city names, municipal names, administrative district codes, etc.; an LSV (Life Style Value) navigation system linked to the database; a calculation unit that performs calculation processing in cooperation with the database and the LSV navigation system; and an input / output unit that inputs and outputs commands and information and data, The LSV navigation system is comprised of an LSV value calculation unit that calculates an LSV value (Life Style Value), a ranking unit that ranks potential customers, a profiling unit that assigns lifestyle raster data to potential customers and estimates them, and a segmentation unit that groups them into towns and districts. The support system comprises: inputting lifestyle cluster data via the input / output unit; a step of importing customer data of the client town / district from the user-owned data via the input / output unit; a step of grouping the store number-customer information file (store CIF) by block, and after name matching, aggregating the age of the store CIF holder and the lifestyle cluster data for each group by the calculation unit; determining a similarity between the lifestyle cluster data by the LSV navigation system and the calculation unit; a step of calculating potential revenue 1 from the number of potential customers for each block, calculated by (total number of households or total population of block) - (number of existing customer accounts after name integration), or (total number of households or total population of block) - (number of existing customers after name integration), using the LSV navigation system and the calculation unit; a step of calculating a risk value by the LSV navigation system and the calculation unit based on purchasing power that is likely to be linked to luck, interests that are likely to be linked to ability, and age that is likely to be linked to personal connections; a step of calculating an LSV value of each customer in all towns and districts by the LSV value generating unit using the formula "LSV value = similarity of lifestyle cluster data x potential profit 1 x (100 - risk value (%))"; A step of setting target products for each age group based on the risk value (%) used to calculate the LSV value by the calculation unit; A step of extracting, by the calculation unit, product contract results for each block based on the calculated LSV value from the user-owned data for the set product; a step of ranking customers with a high record of contracting for products in each block by the ranking unit; a step of determining the level of the contract record for the product for each block by the calculation unit, and calculating the contract prospect rate for potential customers based on the difference between the contract rate for existing customers in the area with the highest contract record for the product and the area with the lowest contract record for the product; a step of determining the number of potential customers expected for each block by using the block with the highest contract rate among the block group to be analyzed as a standard, and multiplying the contract rate of the potential customers by the standard to determine the contract rate of other blocks based on the standard; a step of multiplying the number of potential customers excluding existing customers in a block by the contract prospect rate of the potential customers to obtain a maximum value of the number of potential customers; a step of calculating a total sum of the expected number of potential customers by aggregating all groups in the same age category, aggregating all age categories, and aggregating all products by the calculation unit; a step of calculating a potential profit 2 of the expected number of potential customers by the calculation unit; While carrying out the above, The support system calculates the potential revenue 1 and the potential revenue 2 by (annual income or annual transaction amount of the town / district using statistical information) / (total population of the town / district), or (annual income or annual transaction amount of the town / district using statistical information) / (total number of accounts of the town / district), and is characterized by making it possible to conduct business based on the expected number of potential customers on a block-by-block basis.

2. A method for supporting the discovery of potential customers as described in claim 1, wherein the purchasing power is characterized by family structure, hobbies, real estate, educational background, occupation, income, and credit history; the interests are characterized by family structure, hobbies, investments, educational background, occupation, income, and credit history; and the age is characterized by family structure, hobbies, health, educational background, occupation, income, and credit history.

Images