KR100449083B1 - Method of CRM on comparison with web-log and user ID - Google Patents

Abstract

The present invention relates to a customer relationship management method, and has been proposed to achieve an object of providing an improved customized service by enabling detailed analysis of members as well as information on anonymous users.

In order to achieve the above object, the present invention provides a plurality of visitor clients connected through the Internet, and a website in which a login script for customer authentication is configured on a website for customer relationship management performed by a web server and a mining server. Extracting login information of a visitor accessing a web site configured with a login script by accessing the web server; Storing the extracted log-in information and log data generated in accordance with a user's access to a website as separate files; Filtering and storing the stored log file; Transmitting the stored login information file and the filtered log file to the mining server; Extracting initial access information by using the transmitted login information file and log file; We propose a customer relationship management method through a web log and ID matching, including the step of extracting and storing the website access information of the visitor for a predetermined time from the log file based on the extracted initial access information, and a specific web There is an advantage that can provide customized customer management service by obtaining more reliable user information for each user visiting the site.

Description

How to manage customer relationship through web log and user ID matching {Method of CRM on comparison with web-log and user ID}

The present invention relates to a method that can be used in customer relationship management (CRM), and more particularly, to provide a method for realizing higher quality customer service by extracting data according to individual customers.

The Internet-based Internet business market is expected to expand to 600 billion won in 1999, and by 2005, the market will grow by over 200% annually to reach 2.6 trillion won. In particular, global e-commerce was reported to be worth more than $ 2.6 billion in 1996, $ 7.4 billion in 1997, and $ 23 billion in 1998. IDC (International Data Co), a US market research firm, It is expected to surpass at least $ 220 billion.

As a result of the rapid expansion of users and the development of technology, business owners need a more successful website operation method to secure customers and manage the secured customers. relationship management (CRM).

Customer relationship management (CRM) is a customer management system that automates the customer management process. It is an integrated marketing solution that comprehensively analyzes existing customer information, extracts excellent customers, and advances DB marketing based on various information about them. Integrated customer relationship management that keeps customers long by continually delivering the products and services they want by knowing who our customers are and what they want, thereby maximizing their lifetime value Say process.

In other words, the process of planning, supporting, and evaluating marketing activities based on the characteristics of individual customers by analyzing and integrating customer data and information. These customer relationship management processes usually go through four stages: analysis of customer information for marketing planning and strategy, It can be subdivided into marketing plans through analyzed information, customer contact through various contact channels, and analysis and purification of data generated through customer contact. The same concept applies.

There are various ways to accumulate customer information for customer relationship management on the Internet. Among them, customers can be managed in a three-dimensional manner by managing information such as behaviors that customers move on the website and records of purchases and transactions on the website. There is a need to build a process that understands and organizes the information in real time, and technology foundations such as analytics and data mining have evolved to meet these demands.

Web-log refers to information data about a user accessing a specific website through a web server. The web-log is a record of all operations performed by the user through the web server. The user log data stored as is the page view of the website, the page view of the user, the location and method of access, the page view by time, the number of visitors, and the like.

In general, a user connected to the Internet sends a request to a web server to view a particular web page, and the web server accesses various files related to the web page.

As described above, in a web server connected to a specific web page by a user's request, all operations performed by the user's request are left as data in a predetermined position in the web server. Information about image files, image data, and include files related to web pages is stored in log files.

Such a web log is analyzed using a web analysis tool manufactured by Web Trends Co., Ltd. or Changwon Graphics Co., Ltd. or by using an analysis tool provided by the web server itself. It provides various information such as the number of visitors and visitors, the classification of visitors, the aggregate of visitors' ISPs, the statistics of homepage directory and file, and the analysis by time (month, week, day of the week, day, time).

Hereinafter, a description will be given of an analysis method of a general web log conventionally used as described above.

FIG. 1 is a diagram briefly showing a system configuration for explaining a web analytics method for customer relationship management, which is applied in the prior art, directly or through a modem 5 and an ISP server 6 or a proxy server 7. A plurality of first, second, and third clients (1) (2) (3) connected to the web server 10 is shown representatively.

First of all, various types of data are required for the analysis of the web logs, and in the above configuration, various data stored in the proxy server 7, web log files stored in the web server 10, and web data currently in operation can be obtained. have.

The proxy server 7 is generally known as a firewall, and performs a caching to speed up web surfing of Internet users. The proxy server 7 includes a user client 2 and a web server 10 that the user wants to use. Located in the middle, the web server 10 requests the web server 10 instead of the document requested by the client 2 and then transfers the document to the client 2 again. At this time, the document is also stored in the proxy server 7.

After that, when the user requests the same contents again or when another user requests the same contents, the proxy server 7 delivers the contents stored therein to the user rather than requesting the corresponding site.

When the proxy server 7 is used as described above, a record of a job or the like requested by the web server 10 is recorded in the proxy server 7 rather than a log file of the web server 10, and the proxy server 7 is also recorded. Even if a request is made that is not found in (7), the information of the proxy server 7 is recorded instead of the corresponding user information.

The data that can be obtained for the web analysis in the log file of the web server 10, as described above, the number of visitors and visitors, the classification of visitors, the aggregated visitor's access ISP, the homepage directory and file statistics, hourly (monthly) , Week, day, day, time).

Finally, when looking at the information that can be obtained from the web data, the web data refers to the data or content data of the members running on the actual website, the log file stored in the web server 10 has a file name requested by the user It is recorded.

Looking at the process of performing the web analytics based on the various information obtained as described above are as follows.

Figure 2 is a step-by-step progress diagram illustrating a general web analytics process, based on web logs stored in most web servers such as Apache, IIS, and Netscape.

First, in order to analyze the web log, it is necessary to determine in advance what kind of analysis is necessary for the effective operation of the website (S1). According to the specific decision of the analysis target, the user-specific analysis must be performed according to the characteristics of each site. .

Next, we need to convert the data stored in the log file into data that can be analyzed. In other words, the data stored in the log file is a primitive form of data that has a narrow scope of analysis, and should be converted into a more diverse form of analysis (S2).

After the above process, the log is finally analyzed (S3).

Log analysis can be performed through various algorithms according to the elements to be analyzed, and analysis using data mining algorithms such as association rule search, continuous pattern test, cluster analysis, decision tree, neural network model, etc. is also useful. You can do it.

In the above, the process of general web analytics has been described as described above, and a detailed web analytics method will be described below.

FIG. 3 is a diagram for explaining a general method of analyzing web logs, which sequentially illustrates internal processes performed by the web server 10 and the mining server 20 according to a process.

Here, mining refers to a process of extracting useful information from a large amount of data, and a server that performs data mining is called a mining server.

In the web server 10, a log file that stores data about a connection of a plurality of clients 1, 2, and 3 is stored, and is mainly recorded in units of days.

In this case, it is preferable to use the Combined format for the NCSA (National Center for Supercomputing Applications) and the Extende4d Log Format for the W3C (World Wide Web Consortium).

As described above, the log file stored on the web server 10 in units of days is transmitted to the mining server 20 through FTP (File Transfer Protocol) at a time determined by the web server 10. The mining server In (20), the received log file is loaded into the database 22 at regular intervals, and then loaded from the temporary table to the target table by executing a database program such as PL / SQL. Here, PL / SQL (Procedural Language / SQL) is a language made by adding the code composed of procedural units such as C language to SQL, the language of the database used in Oracle.

However, the general web analytics method described above cannot obtain more detailed user information, and does not provide customized customer management service to each customer but only general services based on the log analysis results. It is true. In addition, since a process of transmitting and analyzing the mining server 20 is performed without refining the log file, a large log file has a disadvantage of requiring a lot of time and a high capacity computer processing capacity for transmission and analysis.

The present invention is to solve the customer service problem as described above, and has the greatest object to provide a customized customer management service by obtaining more reliable user usage information for each user visiting a specific website.

1 is a diagram briefly showing a system configuration for explaining a web analytics method for customer relationship management applied in the prior art

Figure 2 is a step-by-step progress diagram showing a general process of general analytics

3 is a view for explaining a general method of analyzing web logs

4 is a diagram illustrating an internal process of a system for explaining a customer relationship management method according to the present invention.

<Description of the symbols for the main parts of the drawings>

1,2,3: Client 5: Modem

6: ISP server 7: Proxy server

10: web server 11: mining server

In order to achieve the above object, the present invention is a web site 10 and a mining server (Mining server) for the customer relationship management performed by the website 20, the login script for customer authentication is configured on the website, the Extracting login information of a visitor accessing a web site configured with a login script by accessing a web server; Storing the extracted log-in information and log data generated in accordance with a user's access to a website as separate files; Filtering and storing the stored log file; Transmitting the stored login information file and the filtered log file to the mining server; Extracting initial access information by using the transmitted login information file and log file; Based on the extracted initial access information, a web log and a customer relationship management method through ID matching are included, including the step of extracting and storing the website access information of the visitor for a predetermined time from the log file.

The login information includes user information stored in a web program, that is, a CGI (or ASP, PHP, JSP) program, and information generated from a system of a web server. The login information file and filtering of a visitor stored in the web server are also included. The log file further includes transmitting to the mining server through FTP.

Hereinafter, a detailed description of a method used for customer relationship management through weblog ID matching according to the present invention will be described in detail with reference to the accompanying drawings.

4 is a diagram sequentially illustrating an internal operation performed in each device to explain a customer relationship management method according to the present invention, and includes a plurality of clients (1) (2) (3) and a web server (10). And a process performed by the mining server 20.

Looking at the operation performed in the web server 10, first, in order to more detailed log analysis for the individual customer in the present invention, the login information is further used, the login information is visited by a customer to a specific web page member service In order to use the information generated in the process of confirming that the member, that is, information extracted arbitrarily, such as ID, password.

The customer enters his or her ID and password in a web page configured to log in, and the login page inquires the ID and password in the member database where the member information is recorded. The customer confirmed to be a member in the member database is authorized to use the member service.

This login process is performed by a program called CGI (or ASP, PHP, JSP), and uses a program in which a program code for extracting a client's login information is added to the above program. In the description of the operation, in general, the visitor inputs a user ID and password for logging in a website written by a web program such as CGI (or ASP, PHP, JSP). Stored. In addition, the login information of the visitor by combining the date, time, visitor IP address provided by the web server system or web program accessed by the visitor, and browser information and OS information transmitted by the visitor through his web browser. The user information collected at each information provider by this additional program takes the following example format.

alivekim 211.207.52.40 2001-07-09 00:54:25 Mozilla / 4.0 (compatible; MSIE 5.5; Windows 98)

In other words, a member with the name alivekim is logged in through the browser of Mozilla / 4.0 (compatible; MSIE 5.5; Windows 98) at 0:54:25 on July 9, 2001 at 211.207.52.40. I say I confirmed it.

These login information is accumulated in a file every time the customer is confirmed to be a member through the login process, hereinafter this file is called a login information file, and this login information file is a predetermined time separately from the log file stored in the web server 10 It is stored in units (eg daily).

As described above, the login information file is stored with respect to the login of the customer, and the information on the login page requested by the customer is stored in the log file in the web server 10.

Next, the log file stored in the web server 10 is filtered to filter only the data necessary for log analysis. The filtering of the log file uses a scheduler of the web server 10 as a known technology. To run at regular intervals (for example, per day).

The filtering operation refers to a process of removing unnecessary data in order to facilitate processing of a large log file, and whether the log format is a combined format of NCSA or an extended log format of W3C to match log analysis in the following format. To filter.

2001-07-09 01:16:20 210.101.41.91 G /cgi/quiz/quiz1.php3-200--mozilla / 4.0 + (compatible; + MSIE + 5.0; + Windows + 98; + DigExt)

The data above shows that a visitor to a website with an IP address of 210.101.41.91 is running mozilla / 4.0 + (compatible; + MSIE + 5.5; + Windows + 98 at 1:16:20 AM on July 9, 2001. ; + DigExt) requested a /cgi/quiz/quiz1.php3 page using a web browser (Explorer 5.5, using Window98) and sent with GET without any error (what '200' means).

The GET method is one of data transmission methods used in a CGI program. The GET method is mainly used when there are not many input data values or is simply attached to a URL and transmitted to a CGI program.

The log file and the login information file filtered by the web server 10 are transmitted to the mining server 20 through FTP (File Transfer Protocol) at a predetermined time.

Hereinafter, a process performed by the mining server 20 will be described.

The following process is also preferably performed every time determined by the scheduler of the mining server 20, the log file and the log information file filtered by the web server 10 is performed after the FTP transmission to the mining server 20 It is desirable to be.

The mining server 20 which receives the log file filtered through the web server 10 and the user's login information file from the web server 10 reads each file and stores the file in its own memory. At this time, there may be one or more log files.

Thereafter, the mining server 20 finds a website moving path for each visitor by comparing the log data of the same time zone stored in the log file while reading the login information of the stored login information file. The process of finding is as follows.

First of all, the IP address and browser information of the login information file are compared with the IP address and browser information of the log data in the log file stored in the memory, and the customer first visits the log data, and the web page included in the searched log data is found. Becomes the first visited webpage in the customer's website navigation path.

After that, the route is continuously searched for a predetermined time based on the log data searched above. In this case, the route is searched by comparing IP addresses and browser information, similarly to finding the first visit point. Here, a certain time is the average time from the time when a visitor enters a website and receives and closes the service, and usually 30 minutes is set as the average visit time per website, which may vary depending on the administrator's setting. May vary.

For example, the process of finding a website moving path for each customer described above will be described in detail. For example, the login information of the visitor recorded in the login information file may be, for example,

alivekim 211.207.52.40 2001-07-09 00:54:25 mozilla / 4.0 (compatible; MSIE 5.5; Windows 98),

Visitor log data recorded in the log file,

1001: 2001-07-09 00:54:20 210.101.41.91 GET /cgi/quiz/quiz1.php3-200--mozilla / 4.0 (compatible; MSIE + 5.0; Windows + 98; DigExt)

1002: 2001-07-09 00:54:22 211.207.52.40 GET /cgi/quiz/quiz1.php3-200--mozilla / 4.0 ( compatible; MSIE 5.5; Windows 98 )

1003: 2001-07-09 00:54:26 210.101.41.91 GET /cgi/quiz/quiz1.php3-200--mozilla / 4.0 (compatible; MSIE + 5.0; Windows + 98; DigExt)

In the above data, alivekim coincides with the IP address of the log data of line 1002 and the browser information among the log data of 1001,1002,1003, indicating that the log data of line 1002 is one of alivekim's moving paths. Able to know.

In other words, alivekim member can see that the page /cgi/quiz/quiz1.php3 is normally requested (0:54:22 seconds) on July 9, 2001.

The numbers 1001, 1002, and 1003 are line numbers arbitrarily added for description.

The customer information extracted by comparing the filtered log file and the log information file as described above, that is, the website movement path for each customer is stored in the mining server 20 as a file, which is stored in the following preferred form.

alivekim 20010709123436--0 search.naver.com /search.naver-

alivekim 20010709123436 200 G 2 www.ecminer.com /index.html-

alivekim 20010709123438 200 G 58 www.ecminer.com /introduction.html-

In other words, a member with the name alivekim entered the site www.ecminer.com on July 9, 2001, at 12:34:36 AM and searched for it on www.naver.com. I stayed at /index.html for 2 seconds at 12:34:36 AM on 9th and 58 seconds at /introduction.html at 12:34:38 AM on 9th July 2001. It means.

The analysis file in which all movement paths for each customer are stored in the above format is stored in a temporary table by the loader of the DBMS (Oracle, MS-SQL, etc.), and the log analysis and analysis by customer is performed again using a database program called PL / SQL. After the transformation process suitable for mining is stored in the log analysis table of the mining server 20.

The analysis file stored in the log analysis table may be applied to analysis for customer relationship management such as log analysis information and a period database such as a customer (member) database and a purchase (transaction) database operated on a website.

That is, basic log analysis such as visitor and visitor number, visitor's classification, visitor's access ISP's statistics, homepage directory and file statistics, hourly (month, week, day, day, hour) analysis, popular page by member, visit time, It is also possible to analyze members by visit period, movement route, etc. and data mining algorithm such as association rule exploration, continuous pattern inspection, cluster analysis, decision tree, neural network model, etc. together with the period database.

According to the customer relationship management through the web log and ID matching according to the present invention proposed as described above, while the existing log analysis method shows only simple statistical information such as the number of visits, the number of visits, the statistics by date, for anonymous visitors, It has the advantage of enabling detailed analysis of members as well as information on anonymous users.

In addition, since the data mining analysis is linked with the main database, the three-dimensional analysis is performed in the customer relationship management. In addition, since the filtering process is reduced in the web server, the time required for the analysis is reduced. It is a natural effect that the shortening effect is occurring.

Claims (3)

On a website that is connected via the Internet and has a login script for customer authentication on the website for customer relationship management performed by multiple visitor clients, web servers and mining servers, Extracting login information of a visitor accessing a web site configured with a login script by accessing the web server; Storing the extracted log-in information and log data generated in accordance with a user's access to a website as separate files; Filtering and storing the stored log file; Transmitting the stored login information file and the filtered log file to the mining server; Extracting initial access information by using the transmitted login information file and log file; Extracting and storing website access information of a visitor for a predetermined time from the log file based on the extracted initial access information; Customer relationship management method through weblog and ID matching, including The method according to claim 1, The login information includes a web log and ID matching method including using user information stored in a web program such as CGI (or ASP, PHP, JSP) and information generated from a system of a web server. The method according to claim 1, The login information file and the filtered log file of the visitor stored in the web server are further transmitted to the mining server via FTP.