JP2008065500A - Recommend engine installation effect prediction system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provided a recommend engine installation effect prediction system for predicting effects based on recommend engine installation to a Web site. <P>SOLUTION: In installing a recommend engine to a Web site managed by an enterprise AWeb servers 10-1 to enterprise DWeb server 10-4 connected to a network 13, pre-installation sales 63 and pre-installation element data 65 are input from a terminal 9 to a Web server 3. The Web server 3 calculates sales prediction formula prior to the recommend engine installation by applying the input pre-installation sales 63 and pre-installation element data 65 to a statistical model, and calculates sales prediction formula posterior to the recommend engine installation from influence quantity to be given to the pre-installation element data 65 by recommend engine installation. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a system for predicting the effect of introducing a recommendation engine.

  With the spread of communication networks such as the Internet, information distribution services for distributing a wide variety of contents to users have appeared. In such a content distribution system, it is necessary to efficiently distribute optimal information based on user preference in order to propose optimal products and services that meet the user's wishes and requests. Therefore, information on the user's preference is retained, products and services to be provided corresponding to the user's preference are narrowed down, and the information is distributed to the user. (See Patent Document 1 and Patent Document 2)

JP 2004-341784 A Japanese Laid-Open Patent Publication No. 2006-18755

  However, in the above method, it is necessary to modify the existing website and system in order to introduce the recommendation engine or the like into the website or system, and there is a cost for that. However, as a result of introducing the recommendation engine, sales It does not specifically indicate how much has increased. When actually introducing a recommendation engine into an existing system, it is necessary to give the companies considering the introduction of the recommendation engine a decision to consider whether or not to install the recommendation engine, and with the introduction of the recommendation engine. In order to explain the risk, it is necessary to show how much effect can be expected by introducing the recommendation engine.

  The present invention has been made in view of such problems, and an object of the present invention is to provide a recommendation engine introduction effect prediction system and the like that can predict an effect of introduction of a recommendation engine to a website.

  In order to achieve the above-described object, a first invention is a recommendation engine introduction effect prediction system having a server and a database server, and the server inputs purchase history data before introducing the recommendation engine to a website. A purchase history input means before introduction, a calculation means for calculating a sales forecast formula before introducing the recommendation engine on the website using the purchase history data before the introduction, and an input using the sales forecast formula A recommendation engine introduction effect prediction system comprising: a prediction unit that predicts sales after introduction of the recommendation engine based on the determined conditions.

  The server uses the input purchase history data such as the click rate, repeat rate, and sales before the introduction of the recommendation engine, and calculates the sales prediction formula before the introduction of the recommendation engine by applying it to the existing statistical model. When a recommendation engine is introduced, it is affected by various factors and sales increase. Therefore, a sales forecast formula after the recommendation engine is introduced is obtained. By inputting the number of visitors to the website in the sales forecast formula, the sales after the introduction is predicted.

  After introducing the recommendation engine on the website, the server inputs the purchase history data such as click rate, repeat rate, sales, etc. after installing the recommendation engine, and stores it in the database server. It can be reflected. Also, by searching the database server, it is possible to browse as reference information when introducing the recommendation engine.

  The second invention introduces the recommendation engine to the website using the pre-introduction purchase history input means for inputting the purchase history data before the introduction of the recommendation engine to the website and the purchase history data before the introduction. A server comprising: calculation means for calculating a previous sales prediction formula; and prediction means for predicting sales after introduction of a recommendation engine based on an input condition using the sales prediction formula. is there.

  A third invention is a program characterized by causing a computer to function as the server of the second invention.

  ADVANTAGE OF THE INVENTION According to this invention, the recommendation engine introduction effect prediction system etc. which can estimate the effect by recommendation engine introduction to a web site can be provided.

  DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of a recommendation engine introduction effect prediction system and the like according to embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, the configuration of the recommendation engine introduction effect prediction system 1 according to the present embodiment will be described with reference to FIGS. 1, 2, 3, 4, 5, 6, and 7.
FIG. 1 is a block diagram showing a configuration of a recommendation engine introduction effect prediction system 1.

  The recommendation engine introduction effect prediction system 1 includes a Web server 3, a company A Web server 10-1, a company B Web server 10-2, a company C Web server 10-3, a company C Web server 10-4, etc. -1 to 11-N and the like are connected via the network 11. The network 13 is a network such as the Internet.

The Web server 3 is a computer for predicting the effect when a recommendation engine is introduced to a Web site operated by the company A Web server 10-1 to the company D Web server 10-4 connected to the network 13, and the DB server 5 , DB7, terminal 9, etc.
The terminal 9 is a computer for inputting purchase history data and sales before introducing the recommendation engine to the Web site connected to the network 13, purchase history data and sales after introduction of the recommendation, and the like to the Web server 3.
The DB server 5 accumulates data such as purchase history data before the introduction of the recommendation engine on the website, sales, purchase history data after introduction, and sales in the DB 7. Further, the DB server 5 searches the data stored in the DB 7 in response to a request from the Web server 3 and browses information before and after the introduction of the past recommendation engine.

  The company A Web server 10-1 to the company C Web server 10-4 are Web servers that operate the websites of the respective companies. The company A Web server 10-1 to the company C Web server 10-4 have introduced a recommendation engine or are considering introduction.

  The user terminals 11-1 to 11-N are computers connected to the network 13, and browse and access the websites operated by the company A Web server 10-1 to the company D Web server 10-4.

Next, the Web server 3 will be described.
FIG. 2 is a diagram illustrating an example of a hardware configuration of the Web server 3, and FIG. 3 is a diagram illustrating details of the storage device 22 of the Web server 3. The Web server 3 includes a control unit 21, a storage device 22, a media input / output unit 23, an input unit 25, a printing unit 27, a display unit 26, and a communication unit 27 that are interconnected by a bus 28.

  The control unit 21 includes a CPU (Central Processing Unit) that executes a program, and a memory such as a ROM (Read Only Memory) and a RAM (Random Access Memory) that store program instructions or data. The control unit 21 controls the operation of the entire Web server 3.

The storage device 22 is a storage medium for storing fixed data such as a control program of the Web server 3, an image file, and the like.
As shown in FIG. 3, the storage device 22 includes a control program such as the OS 31, a recommendation engine introduction effect prediction system 33 for predicting the effect of introduction of the recommendation engine to the website, and the like.

  The media input / output unit 23 is a medium drive such as CD-ROM (Compact Disc Read Only Memory) or CD-RW (CD-ReWriteable), a flexible disk, and MO (Magneto Optical Disc). Read and write data to the medium.

The input unit 24 is an input device such as a keyboard and a mouse.
The printing unit 25 is a printer, and prints necessary information and the like according to a request from the user.
The display unit 26 is a display device such as a CRT (Cathode-Ray Tube) or an LCD (Liquid Crystal Display).
The communication unit 27 is a communication control device, a communication port, and the like, and controls communication via the network 13, communication with the DB server 5, and the like.

Next, the DB server 5 will be described.
FIG. 4 is a diagram illustrating an example of a hardware configuration of the DB server 5, and FIG. 5 is a diagram illustrating details of the storage device 42 of the DB server 5.

  As shown in FIG. 4, the DB server 5 includes a control unit 41, a storage device 42, a media input / output unit 43, an input unit 44, a printing unit 45, a display unit 46, and a communication unit 47 interconnected by a bus 48. .

  The control unit 41 includes a CPU that executes a program and a memory such as a ROM and a RAM that store program instructions or data. The control unit 41 controls the operation of the entire DB server 5.

The storage device 42 is a storage medium for storing fixed data such as a control program of the DB server 5.
As illustrated in FIG. 5, the storage device 42 includes a control program such as the OS 51, a data management program 53 for managing data stored in the DB 7, and the like.

  The media input / output unit 43 is a drive of a medium such as a CD-ROM or CD-RW, flexible disk, or MO, and reads data from the medium or writes data to the medium.

The input unit 44 is an input device such as a keyboard and a mouse.
The printing unit 45 is a printer, and prints necessary information and the like according to a request from the user.
The display unit 46 is a display device such as a CRT or LCD.
The communication unit 47 is a communication control device, a communication port, and the like, and controls communication with the Web server 3 and the like.

Next, DB7 will be described.
FIG. 6 is a diagram showing details of the DB 7, and FIG. 7 is a diagram showing details of the company data 61 of the DB 7.

  As shown in FIG. 6, the DB 7 has company A data 61-1, company B data 61-2, company C data 61-3, and company D data related to companies that operate various websites connected to the network 13. 61-4 etc.

As shown in FIG. 7, the company data 61 includes pre-introduction sales 63, pre-introduction element data 65, post-introduction sales, post-introduction element data 69, and the like.
The pre-introduction sales 63 are daily and monthly sales before the recommendation engine is introduced to the company's Web site.
The post-introduction sales 65 is the daily or monthly sales after the recommendation engine is introduced to the website.

  The pre-introduction element data 65 and the post-introduction element data 69 are element data serving as various parameters before and after the introduction of the recommendation engine to the website, respectively. Click rate 71, repeat rate 72, purchase item number 73, Site design 74, handling product price 75, handling product number 76 and so on.

The click rate 71 is the ratio of those who have advanced to the detailed screen of the product or service among all visitors to the Web site.
The repeat rate 72 is a ratio of people who visit the website more than once out of all visitors to the website.
The number of purchased items 73 is the number of items of products or services purchased in one purchase process on the website.
The site design 74 indicates which level of the website the recommendation engine is in. The first level is the top page of the website, the second level is the page displayed after the top page of the website, The third level represents a page displayed next to the second level of the website.
The handling product price 75 is a product or service average price handled in the Web site.
The number of handled products 76 is the number of products or services handled in the website.

Next, processing for predicting the effect when the recommendation engine is introduced to the website by the recommendation engine effect prediction system 1 will be described.
FIG. 8 is a graph showing the effect of introducing the recommendation engine to the website, and FIG. 9 is a diagram showing an example of the recommendation engine introduction effect prediction display screen 900 on the Web server 3.

When a recommendation engine is introduced into a website operated by the company A web server 10-1 to the company D web server 10-4, etc., sales before introduction 63 are introduced as data before introduction of the recommendation engine from the terminal 9 connected to the web server 3. The pre-introduction element data 65 is input.
The Web server 3 uses the input pre-introduction sales 63 and pre-introduction element data 65 to apply a regression model, which is an existing statistical model, so that the pre-introduction sales y1 = ax + b shown in FIG. calculate.
By calculating the sales y1 = ax + b before introduction of the recommendation engine, the angle θ1 representing the sales forecast amount with respect to the number of visitors is determined from the formula θ1 = tan ⁻¹ ((y1−b) / x).

When the recommendation engine is introduced to the website, the data shown in the pre-introduction element data 65 is affected, and the sales on the website increase accordingly. The value of element data that affects sales varies depending on the characteristics of the website. For example, it is assumed that the click rate 71 increases by 20% by introducing a recommendation engine in the website. When the click rate 71 before introduction is 10%, the click rate 71 after introduction is 12%, and when the click rate 71 before introduction is 30%, the click rate 71 after introduction is 36%.
Assuming that the repeat rate 72 increases by 10% due to the introduction of the recommendation engine, when the repeat rate 72 before introduction is 10%, the repeat rate 72 after introduction is 11%, and the repeat rate 72 before introduction is 30% The repeat rate 72 after introduction is 33%.
Assuming that the number of purchased items 73 increases by 10% due to the introduction of the recommendation engine, when the number of purchased items 73 before introduction is 10, the number of purchased items 73 after introduction is 11, and the number of purchased items 73 before introduction is In the case of 100, the number of purchased items 73 after introduction is 110.

The click rate 71 changes before and after the introduction due to the site design 74 that indicates the hierarchy of the recommendation engine in the website. For example, after the recommendation engine is introduced, when the site design 74 is the first hierarchy, the click rate 71 is increased by 20% before and after the introduction, the click rate 71 is increased by 10% in the second hierarchy, and the click rate is increased in the case of the third hierarchy. 71 is assumed to increase by 5%.
When the site design 74 is the first level, if the click rate 71 before introduction is 10%, it will be 12% after introduction. In the case of the second level, if the click rate 71 before introduction is 10%, it will be 11% after introduction. In the case of the third hierarchy, if the click rate 71 before introduction is 10%, it increases to 10.5% after introduction.

The click rate 71 changes depending on the handling product price 75 and the handling product number 76.
For example, it is assumed that the click rate 71 changes by 10% in proportion to the handling product price 75. If the probability of clicking before introduction is 50%, the click rate 71 after introduction is 50 × (1 + 0.1) = 55% when the handling product price 75 is 1000 yen, and the handling product price 75 is 2000 yen. In this case, 50 × (1 + 0.2) = 60%.
It is assumed that the click rate 71 changes by 10% in proportion to the number of products handled 76. Assuming that half of the number of handled products 76 is clicked, when the number of handled products 76 is 100, 50 products are clicked before introduction, and 55 products are clicked after introduction. Assuming that the number of handled products 76 is 200, the click rate is assumed to increase by 20%. Therefore, 100 products are clicked before introduction, and 120 products are clicked after introduction.

Thus, the click rate 71, repeat rate 72, purchase item number 73, site design 74, handling product price 75, handling product number 76 of the pre-introduction element data 65 are affected by the introduction of the recommendation engine, α1 to α6, respectively. Then, the angle θ2 representing the sales forecast amount with respect to the number of customers after introduction of the recommendation engine can be defined by the equation θ2 = tan ⁻¹ ((y1−b + Σαn) / x).
As a result, as shown in FIG. 8, the sales prediction formula y2 = cx + b when the recommendation engine is introduced can be obtained.

The Web server 3 displays a recommendation engine introduction effect prediction display screen 900 as shown in FIG. 9 on the display unit 26 based on the calculated sales forecast amount for the number of visitors before and after the introduction of the recommended engine.
The recommendation engine introduction effect prediction display screen 900 includes a graph 901 showing the introduction effect of the recommendation engine shown in FIG. 8, an expected number of customers input field 902, an execution button 903, and a recommendation engine introduction effect expected amount 904.
When the user inputs the expected number of visitors in the expected number of visitors input field 902 and clicks the execution button 903 with the input unit 24 such as a mouse, the control unit 21 of the Web server 3 uses the input number of expected visitors as a sales prediction formula y2. By substituting into = cx + b, the expected amount is calculated and displayed in the recommended engine introduction effect expected amount 904. The user roughly grasps the effect of the recommendation engine introduction with reference to the recommended engine introduction effect expected amount 904 displayed, and the website where the recommendation engine is actually operated by the company AWeb server 10-1 to the company DWeb server 10-4 Consider whether to introduce to.

  As described above, according to the present embodiment, a recommendation engine introduction effect prediction system that can predict the effect of introduction of a recommendation engine on a website can be provided.

In addition, post-installation sales 67 after introducing the recommendation engine on the website, introduction of click rate 71, repeat rate 72, number of purchased items 73, site design 74, handling product price 75, handling product number 76, etc. By inputting the post-element data 69 from the terminal 9 to the Web site 3 and accumulating it in the DB 7 via the DB server 5, it is possible to search for cases of introducing the recommendation engine.
Furthermore, using the input post-introduction sales 67 and post-introduction element data 69, the influence amount of various element data before and after the introduction of the recommendation engine according to the characteristics of the website is corrected to calculate the sales forecast formula y2 = cx + b. By reflecting it, it is possible to predict the recommendation engine introduction effect with higher accuracy.

  The preferred embodiment of the recommendation engine introduction effect prediction system according to the present invention has been described above with reference to the accompanying drawings, but is not limited to the above-described embodiment. It is obvious for those skilled in the art that various modifications or modifications can be conceived within the scope of the technical idea described in the claims, and these are naturally within the technical scope of the present invention. It is understood that it belongs.

The block diagram which shows the structure of the recommendation engine introduction effect prediction system 1 The figure which shows an example of the hardware constitutions of the Web server 3 The figure which shows the detail of the memory | storage device 22 of the Web server 3 The figure which shows an example of the hardware constitutions of DB server 5 The figure which shows the detail of the memory | storage device 42 of DB server 5 Diagram showing details of DB7 The figure which shows the details of company data 61 of DB7 Graph showing the effect of introducing the recommendation engine to the website The figure which shows an example of the recommendation engine introduction effect prediction display screen 900 on the Web server 3

Explanation of symbols

1 ……… Recommendation Engine Introduction Effect Prediction System 3 ……… Web Server 5 ……… DB Server 7 ……… DB
9 ……… Terminal 11 ……… User terminal

Claims (8)

A recommendation engine introduction effect prediction system having a server and a database server,
The server
A pre-installation purchase history input means for inputting purchase history data before introducing the recommendation engine on the website;
Using the purchase history data before introduction, calculating means for calculating a sales forecast formula before introducing the recommendation engine on the website;
Predicting means for predicting sales after the introduction of the recommendation engine based on the input conditions using the sales prediction formula;
A recommendation engine introduction effect prediction system comprising: The server
After-introduction purchase history data input means for inputting purchase history data after introducing a recommendation engine on the website;
Sales forecast formula reflecting means for reflecting the purchase history data after the introduction to the sales forecast formula;
The recommendation engine introduction effect system according to claim 1, further comprising:   The recommendation engine introduction effect prediction system according to claim 1, wherein the prediction means predicts sales after introduction of the recommendation engine based on the number of visitors to the website.   The recommendation engine introduction effect prediction system according to claim 1, wherein the pre-introduction purchase history input unit inputs a click rate, a repeat rate, and sales on the website before introducing the recommendation engine.   The recommendation engine introduction effect prediction system according to claim 1, wherein the post-introduction purchase history input unit inputs a click rate, a repeat rate, and sales on the website after the introduction of the recommendation engine. A pre-installation purchase history input means for inputting purchase history data before introducing the recommendation engine on the website;
Using the purchase history data before introduction, calculating means for calculating a sales forecast formula before introducing the recommendation engine on the website;
Predicting means for predicting sales after the introduction of the recommendation engine based on the input conditions using the sales prediction formula;
A server comprising: After-introduction purchase history data input means for inputting purchase history data after introducing a recommendation engine on the website;
Sales forecast formula reflecting means for reflecting the purchase history data after the introduction to the sales forecast formula;
The server according to claim 6, further comprising:   A program causing a computer to function as the server according to claim 6.

Images