JP2017220155A - Customer management in electronic commerce site - Google Patents

Abstract

PROBLEM TO BE SOLVED: To monitor/analyze temporal changes of the purchasing behavior of each customer in an electronic store to make it possible to conduct individual sales promotion activity to each customer at appropriate timing.SOLUTION: Purchase records are periodically totaled, and cumulative data for each customer including the amount of purchase, the number of purchases and final purchase date in the latest prescribed period is obtained for each customer. Processing of ranking the degree of excellence of a customer is executed periodically on the basis of the cumulative data for each customer. In the ranking processing, a rank correlated with a customer ID is determined by an algorithm using a large amount of purchase as a rank-up factor, a large number of purchases as a rank-up factor, and a small number of days from the final purchase date to a given day as a rank-up factor.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a computing technology for analyzing a customer's purchasing behavior at an electronic commerce site (sometimes referred to as an electronic store) that sells various products (goods, services, real estate, etc.) to the customer through the Internet. The present invention relates to an analysis method focusing on time-series changes in purchasing behavior for each customer.

  As is well known, many retailers build long-term good relationships by using a computing system to record and manage customer attributes and contact histories, and respond carefully to each customer. , We are working to improve customer satisfaction. Such an approach is called CRM (Customer Relationship Management).

  In a well-known CRM, for example, a database is used to record and manage detailed attribute information and purchase history of each customer, the contents of inquiries and complaints, etc., and promptly and appropriately respond to inquiries. Mainly activities such as introducing products that suit the needs. By maintaining good relationships with individual customers, you can expect them to give priority to consideration over other companies in the next replacement, additional purchase, purchase of other products, etc. CRM is being implemented for the purpose of evaluating the company (product) and improving the image in various surveys.

=== Objectives of the Invention ===
An object of the present invention is to monitor and analyze temporal changes in purchasing behavior of individual customers in an electronic store, and to make it possible to carry out individual promotional activities for individual customers at an appropriate timing.

=== Technical matters at the heart of the invention ===
The present invention created to achieve the above object can be basically grasped as a method specified by the following items (1) to (5) (this is referred to as invention 1).
(1) It is a computing method implemented in the computing system of an electronic commerce site. (2) This system uses a customer ID, a purchase price, and a purchase date that are generated each time a customer makes a purchase at an electronic store. (3) This system periodically aggregates the stored purchase records, and totals by customer including the purchase price, the number of purchases, and the last purchase date for the most recent specified period for each customer. Obtaining data (4) This system periodically executes a process of ranking customer excellence based on aggregated data by customer (5) The ranking process ranks up that the purchase amount is large An element with a higher number of purchases as a rank-up element and an element with a lower number of days from the last purchase date to the current day as a rank-up element By Gorizumu, to determine the rank and association with the customer ID

Invention 2 is a method specified by the following item (6) on the premise of Invention 1.
(6) When outputting the rank for each customer, this system compares the current rank with the previous rank for each customer, extracts the customers that have fallen below the standard, and outputs an alert.

Invention 3 is a method specified by the following matter (7) on the basis of Invention 1 or 2.
(7) When outputting the rank for each customer, this system compares the current rank with the previous rank for each customer, extracts the customers who have been ranked higher than the standard, and outputs an alert.

Invention 4 is a method specified by the following matter (8) on the premise of any one of Inventions 1 to 3.
(8) The system outputs the number of customers by rank as it outputs the rank by customer.

Invention 5 is a method specified by the following items (9) and (10) on the premise of any one of Inventions 1 to 4.
(9) The system classifies customers into a plurality of clusters by paying attention to a predetermined attribute. (10) The system outputs the rank for each customer and the number of customers for each rank and each cluster. Output

The invention 6 is a method specified by the following items (11) and (12) on the basis of the invention 5.
(11) This system can pre-set sales promotion measures to be implemented for customers who have been ranked up or down above the standard for each cluster. (12) This system can be ranked up or down above the standard. When the selected customers are extracted, the pre-set sales promotion measures for each cluster are automatically executed.

The invention 7 is a method specified by the following items (13) to (15) on the basis of the invention 5.
(13) This system sets the period sales target and monitors the progress rate during the period. (14) This system performs the sales promotion measures to be implemented for customers who rank down beyond the standard by rank and by cluster. (15) When the progress rate during the period to be monitored does not reach the standard, this system can be used for customers who rank down above the standard and for each rank and cluster. Automatically implement sales promotion measures

Invention 8 is a method specified by the following items (21) to (23).
(21) It is a computing method implemented in a computing system in an online shopping mall. (22) This system is configured to store a customer ID and a retail shop ID each time a customer makes a purchase at a retail shop in an electronic shopping mall. A purchase record including at least a purchase amount and a purchase date is generated, the generated purchase record is aggregated in the mall management module, and the purchase record is transmitted to the retail store module corresponding to the retail store ID of the generated purchase record ( 23) The retail store module of each retail store performs the computing method according to any one of inventions 1 to 7 on the received purchase record of the own store.

=== Effect of Invention ===
According to the present invention, it is possible to understand temporal changes in purchase behavior of individual customers in an electronic store in an easy-to-understand format called ranking, thereby making it easy to conduct individual promotional activities for individual customers at an appropriate timing. Can be implemented.

Explanatory diagram of customer segment classification in the embodiment

=== Purchase records for each customer regularly and rank ===
As is well known, in an electronic store, every time a customer purchases a certain product, detailed information on the purchase content is recorded in association with the customer ID. This is a purchase record of a purchase behavior unit. The online store computing system includes a CRM system according to the present invention. The CRM system aggregates purchase records by customer as described in detail below.

In one embodiment, a totaling process is performed once a month, and for each customer, customer-specific total data consisting of the following three items associated with a customer ID is obtained based on the most recent one year purchase record.
● Purchase amount (total amount for the past year)
● Number of purchases (total for the past year)
● Last purchase date (recent purchase date in the past year)

  In the aggregation process of one embodiment, the customer's superiority is ranked by the algorithm described below based on the above-mentioned aggregate data for each customer. In the ranking process, first, the purchase amount, the number of purchases, and the last purchase date are scored as follows.

《Scoring of purchase amount》
More than 100,000 yen ... 100 points
5-100,000 yen ... 80 points
3 to 50,000 yen ... 60 points
Less than 30,000 yen ... 40 points

《Score purchase times》
10 times or more: 70 points
5-9 times ... 50 points
2 to 4 times ... 30 points
1 time …… 10 points

《Points for last purchase date》
Within one month ... 40 points
2 to 3 months ago 30 points
4-6 months ago ... 20 points
More than 7 months ago ... 10

The scores of the above three items are totaled, and the ranking is made into the four ranks A to D as follows according to the total score.
150 points or more: Rank A
100-140 points ... rank B
60-90 points ... rank C
50 points or less ... rank D

=== Detecting rank variation ===
The CRM system performs monthly total ranking processing, determines a rank for each customer, and outputs the list. Further, for each customer, the latest rank is compared with the rank given by the aggregate ranking process of the previous month, and rank fluctuation is detected as follows.
《Rank down》
A → B A → C A → D
B → C B → D
C → D
"rank up"
D → C D → B D → A
C → B C → A
B → A

  The CRM system outputs the number of customers for each type of rank variation. Furthermore, some of the above 12 types of rank variation are set as the target types, and for example, customer information corresponding to each of the rank-down types (A → C) and (A → D) is obtained from the customer database. Extract and output alert. It can be freely set in advance which type of customer is the target of alert output.

  By executing the monthly counting process in the CRM system as described above, it is possible to understand at a glance the degree of liveliness / recession of the electronic store from the change in the monthly counting results. It is also possible to understand the relationship and the degree of effectiveness of sales promotions such as advertisements and discount campaigns conducted by electronic store operators. Based on this understanding, the following sales promotion activities can be planned in a timely and appropriate manner. In addition, it is possible to easily implement personal sales promotion measures according to the purchasing behavior of each customer.

=== Cluster classification based on customer attributes and preferences ===
As is commonly done as a CRM activity, there are many items based on personal attributes such as the customer's age, gender, occupation, address, and annual income, and preference information estimated from behavioral history such as purchase history and product browsing history at electronic stores. It is desirable to classify customers into several types (clusters) and perform CRM activities with different contents for each cluster.

  It is desirable to apply cluster classification based on customer attributes and preferences also in the above-described ranking processing based on the aggregated data for each customer, ranking change detection, and alert output according to the present invention. In one embodiment, as shown in FIG. 1, the customer is classified into four clusters 1 to 4 based on personal attributes and preferences, and the customer's superiority is expressed in four stages for each of the clusters 1 to 4. Are classified according to ranks A to D. Thereby, as shown in FIG. 1, the customer decided to classify into 16 types of segments A1 to A4, B1 to B4, C1 to C4, and D1 to D4. This segment classification is the classification by rank and by cluster described above.

  The CRM system outputs the corresponding number of people for each of the 16 types of segments shown in FIG. 1 by the above-mentioned monthly total ranking process. In the CRM system, for example, if a preset (program) is made such as “extract customers ranked down from segment A3 to segment C3 and output alerts”, alerts will be given to the corresponding customers by monthly total ranking processing. Output is generated.

=== Automatic implementation of sales promotion measures based on ranking ===
As can be seen in many CRM systems, an automatic implementation system for a sales promotion measure such as sending an e-mail with contents prepared in advance to a customer corresponding to a preset condition can be incorporated in the present invention. .

  For example, it may be programmed to “send mail X to a customer who has been downgraded from A to B” or “send item Y to a customer who has been ranked from segment C2 to segment A2”. it can. The CRM system is configured so that this program is automatically executed based on the result of the ranking process.

  In addition, the sales target and the progress rate can be added to the conditions for automatically implementing the sales promotion measure. As is done by many retailers, a period sales target for an electronic store is set, and the progress rate is monitored by sequentially comparing the sales amount and the target from the beginning. When the progress rate is below a predetermined standard, the CRM system automatically activates a pre-programmed sales promotion measure. For example, a program “send mail Z to a customer who has been ranked down from segment A1 to segment B1” is executed.

In the embodiment described in detail above, if a certain customer M does not shop for more than one year, the customer M is excluded from the ranking analysis target according to the present invention. Such an event is called rank loss, and customer M is called rank lost customer. As a matter of course, as a desirable CRM activity, a sales promotion measure is separately taken for customers who have lost rank. To do so, list rank-loss customers, classify rank-loss customers according to the length of time since rank loss, generate alert outputs as appropriate, and program for rank-loss customers that have been programmed appropriately. Configure to automatically implement promotional measures. At that time, it is desirable to program sales promotion measures for each cluster, program for each past rank, or program for each rank loss period. That is, it is good also as invention specified by the following matters (16) and (17) on the assumption of the above-mentioned inventions 1-7.
(16) This system manages customers who do not have purchase records in the specified period as rank lost customers and classifies them according to the elapsed period from rank loss. (17) This system meets specified conditions. Automatically implement pre-set sales promotion measures for rank-losing customers

  In addition to email, sales promotion measures that can be implemented automatically include retargeting advertisements, PUSH distribution to advertiser apps, POPUPs when viewing advertiser sites, and specified measures for the user when viewing advertiser sites. It is possible to appropriately adopt an optimized display including

=== Application of the Present Invention to an Electronic Shopping Street ===
As is well known, in an online shopping mall, a retailer module of a retailer who is a tenant is operating on a computing system operated by an operator corresponding to a landlord. Since the retail store module of each retail store receives the purchase records of the retail store from the electronic shopping mall computing system one by one, the above-described computing method of the present invention can be implemented in each retail store module. .

Claims (9)

The method specified by the following items (1) to (5).
(1) It is a computing method implemented in the computing system of an electronic commerce site. (2) This system uses a customer ID, a purchase price, and a purchase date that are generated each time a customer makes a purchase at an electronic store. (3) This system periodically aggregates the stored purchase records, and totals by customer including the purchase price, the number of purchases, and the last purchase date for the most recent specified period for each customer. Obtaining data (4) This system periodically executes a process of ranking customer excellence based on aggregated data by customer (5) The ranking process ranks up that the purchase amount is large An element with a higher number of purchases as a rank-up element and an element with a lower number of days from the last purchase date to the current day as a rank-up element By Gorizumu, to determine the rank and association with the customer ID 2. The method according to claim 1, which is specified by the following item (6).
(6) When outputting the rank for each customer, this system compares the current rank with the previous rank for each customer, extracts the customers that have fallen below the standard, and outputs an alert. The method according to claim 1 or 2, which is specified by the following item (7).
(7) When outputting the rank for each customer, this system compares the current rank with the previous rank for each customer, extracts the customers who have been ranked higher than the standard, and outputs an alert. The method according to any one of claims 1 to 3, which is specified by the following item (8).
(8) The system outputs the number of customers by rank as it outputs the rank by customer. The method according to any one of claims 1 to 4, which is specified by the following items (9) and (10).
(9) The system classifies customers into a plurality of clusters by paying attention to a predetermined attribute. (10) The system outputs the rank for each customer and the number of customers for each rank and each cluster. Output The method according to claim 5, which is specified by the following items (11) and (12).
(11) This system can pre-set sales promotion measures to be implemented for customers who have been ranked up or down above the standard for each cluster. (12) This system can be ranked up or down above the standard. When the selected customers are extracted, the pre-set sales promotion measures for each cluster are automatically executed. The method according to claim 5, which is specified by the following items (13) to (15).
(13) This system sets the period sales target and monitors the progress rate during the period. (14) This system performs the sales promotion measures to be implemented for customers who rank down beyond the standard by rank and by cluster. (15) When the progress rate during the period to be monitored does not reach the standard, this system can be used for customers who rank down above the standard and for each rank and cluster. Automatically implement sales promotion measures The method according to any one of claims 1 to 7, which is specified by the following items (16) and (17).
(16) This system manages customers who do not have purchase records in the specified period as rank lost customers and classifies them according to the elapsed period from rank loss. (17) This system meets specified conditions. Automatically implement pre-set sales promotion measures for rank-losing customers The method specified by the following items (21) to (23).
(21) It is a computing method implemented in a computing system in an online shopping mall. (22) This system is configured to store a customer ID and a retail shop ID each time a customer makes a purchase at a retail shop in an electronic shopping mall. A purchase record including at least a purchase amount and a purchase date is generated, the generated purchase record is aggregated in the mall management module, and the purchase record is transmitted to the retail store module corresponding to the retail store ID of the generated purchase record ( 23) The retail store module of each retail store performs the computing method according to any one of claims 1 to 8 on the received purchase record of the own store.

Images