JP2011039909A - Method and system for optimizing presentation information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and system for optimizing information presented via a computer network. <P>SOLUTION: A first site information presentation space is provided on a display screen of a second site, information is selected from a plurality of pieces of information of the first site, and the selected information is presented to the first site information presentation space on the second site. A behavior history relating to information on the first site information presentation space of a plurality of viewers of the second site is acquired, a recommendation rule particular to the second site is created by the behavior history, and the first site information presented to the first site information presentation space is selected and updated, based on the recommendation rule particular to the second site. A behavior history relating to information on the first site information presentation space of a specific viewer in the second site is acquired, a recommendation rule particular to the specific viewer is created by the behavior history, and the first site information to be presented to the first site information presentation space is updated based on the recommendation rule particular to the specific reader. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

The present invention relates to a method and system for optimizing presentation information.

Information on media such as television, newspapers, and radio is “mass distribution” that is not optimized at all for viewers, and its contents often have nothing to do with the interests of viewers. The same applies to the information posted on the Internet media, and there is little effort to change the content according to the interests of the viewer, which is inefficient. This means that inefficient advertisement distribution is performed when an advertisement is taken as an example.

For example, conventional affiliate advertisements are distributed toward the mass without being optimized, or are simply matched with the text of the content or rough segmentation by site attributes. “Content match advertising” is a method of analyzing texts written in advertising media and delivering advertisements that are strongly related to the texts. The level of optimization is not high in that it does not necessarily match. “Site segment advertisement” is a technique in which attribute data of a user who visits an advertising medium is grasped to some extent, and an advertisement presumed to be interested in the segment is posted. For example, a rule such as “post a cosmetic advertisement on a site with many women in their twenties” is applied. However, even if it is a viewer who is applied to a preset segment, it is not necessarily interested in the advertisement, and the advertisement that can be delivered is limited because segmentation other than the attribute that is grasped in advance cannot be performed There is a drawback.

Thus, information (product advertisements, news, etc.) presented on the Internet media is not optimized for visitors.

Also, Patent Document 1 and Patent Document 2 describe advertising information distribution technology that takes into account user attribute information and action history, but since user attribute information and action history are assumed, the first time user It is not possible to present more optimized advertising information.

JP 2002-203119 A JP2007-156930

The present invention provides a method and system for optimizing information presented via a computer network.

One more specific object of the present invention is to provide a new advertisement distribution method and system in which distribution contents are optimized according to the interests of the viewer in order to eliminate the inefficiency of existing advertisement distribution. There is.

The technical means adopted by the present invention are:
A first site information presentation space for presenting information of the first site providing a plurality of information is provided on the display screen of the second site, selecting information from the plurality of information of the first site, and selecting the selected information Presenting in the first site information presentation space on the second site;
The action history regarding the information on the first site information presentation space of a plurality of viewers of the second site is acquired, and a recommendation rule specific to the second site is created using the action history, and the recommendation rule specific to the second site Selecting and updating the first site information presented in the first site information presentation space on the second site based on
An action history related to information on the first site information presentation space of a specific viewer of the second site is acquired, and a recommendation rule specific to the specific viewer is created using this action history, and a specific recommendation specific to the viewer Selecting and updating the first site information presented in the first site information presentation space on the second site based on the rules;
A method of optimizing presentation information using a computer network equipped with
It is.

In one aspect,
The first site information is linked to the first site and presented in the first site information presentation space on the second site,
In addition to the action history related to information on the first site information presentation space of a plurality of viewers of the second site, the action history related to information on the first site via the second site is acquired, and these action histories are obtained. To create recommendation rules specific to the second site,
In addition to the action history related to the information on the first site information presentation space of the specific viewer of the second site, the action history related to the information on the first site via the second site is acquired, and these action histories are Used to create a specific viewer-specific recommendation rule.

In one aspect,
If the viewer of the second site is the first visitor, present the information selected based on the recommendation rules specific to the second site in the first site information presentation space,
When the viewer of the second site is a repeater, the information selected based on the recommendation rule unique to the viewer is presented in the first site information presentation space.

In one aspect,
When the viewer of the second site is the first visitor, when the viewer selects information on the first site information presentation space, the second site visitor changes to the first information presentation space based on the recommendation rule unique to the second site. Select and present the first site information to be presented in real time,
When the viewer of the second site is a repeater, when the viewer selects information on the first site information presentation space, it is presented in the first information presentation space based on a recommendation rule unique to the viewer. The first site information is selected and presented in real time.

In one aspect,
The creation of the recommendation rule specific to the second site includes obtaining a correlation coefficient between information based on the action history of a plurality of viewers of the second site,
The creation of the recommendation rule specific to the specific viewer includes obtaining a correlation coefficient between information based on the behavior history of the specific viewer of the second site.

In one aspect,
The creation of the recommendation rule specific to the second site includes obtaining a ranking of information that is easily selected on the second site based on the action history of a plurality of viewers on the second site,
Creation of the recommendation rule specific to the specific viewer includes obtaining a ranking of information that can be easily selected by the viewer based on the behavior history of the specific viewer of the second site,
When the visitor of the second site is the first visitor, at the time of visiting the second site, the information selected based on the ranking of information that can be easily selected on the second site is presented in the first site information presentation space,
When the viewer of the second site is a repeater, the information selected based on the ranking of information that can be easily selected by the viewer is presented in the first site information presentation space when visiting the second site.

  In one aspect, in creating the recommendation rule specific to the specific viewer, the recommendation rule specific to the second site is used as the initial recommendation rule, and information on the first site information presentation space of the specific viewer on the second site is used. The recommendation rule specific to a specific viewer of the second site is updated by updating the recommendation rule specific to the second site using the action history related to the information on the first site via the second site. You may create it.

In one aspect, the recommendation rule of the 1st site is created using the action history of a plurality of visitors of the 1st site.
In creating a recommendation rule specific to the second site and a recommendation rule specific to a specific viewer, an action history other than the behavior history of the viewer of the second site or an action history other than the behavior history of the specific viewer (for example, The behavior history of a plurality of viewers of the first site may be additionally used.
In one aspect, in creating a recommendation rule specific to the second site,
Using the recommendation rule of the first site as the initial recommendation rule, action history regarding information on the first site information presentation space of a plurality of viewers of the second site, or information on the first site via the second site The recommendation rule specific to the second site may be created by updating the recommendation rule of the first site using the action history regarding the second site.
In one aspect, in creating a recommendation rule specific to a specific viewer of the second site, the recommendation rule of the first site is used as an initial recommendation rule, and the first site information presentation space of the specific viewer of the second site is used. An action history related to the information on the first site, or an action history related to the information on the first site via the second site, to update the recommendation rules on the first site, thereby recommending rules specific to a specific viewer on the second site May be created.
In one aspect, an action history related to information on the first site via the second site, or an action history related to information on the first site information presentation space of a plurality of viewers of the second site, It may be regarded as an action history on the website and used to create a recommendation rule for the first site.

Other technical means adopted by the present invention are:
A first site providing a plurality of information;
A second site provided on the display screen with a first site information presentation space for presenting information of the first site;
An intermediary server;
The first site, the second site, and the mediation server are connected to each other via a computer network so that data can be transmitted and received.
The intermediary server includes a behavior history storage unit that stores a viewer's behavior history in association with each viewer, a recommendation rule creation unit that creates a recommendation rule for selecting information recommended to the viewer using the behavior history, A recommendation rule storage unit for storing the created recommendation rule, a recommendation information selection unit for selecting recommendation information using the recommendation rule, and the information on the first site in the first site information presentation space of the second site A presentation information storage unit for storing information related to information on the first site, and information presentation means for presenting information selected by the recommendation information selection unit from the presentation information storage unit to the information presentation space on the second site;
With
The behavior history storage unit obtains behavior histories related to information on the first site information presentation space of the viewer of the second site in association with each viewer, so that the first of a plurality of viewers of the second site can be obtained. An action history related to information on the site information presentation space is stored in association with each viewer.
The recommendation rule creation unit
Means for creating a recommendation rule specific to the second site based on an action history related to information on the information presentation space of a plurality of viewers of the second site;
Means for creating a recommendation rule specific to a specific viewer based on an action history related to information on the information presentation space of the specific viewer of the second site;
With
The recommended information selection unit and the information presentation unit are:
Select and update the first site information presented in the first site information presentation space on the second site based on the recommendation rules specific to the second site,
A presentation information optimization system that selects and updates first site information presented in a first site information presentation space on a second site based on a recommendation rule specific to a specific viewer.

In one aspect,
The first site is provided with behavior history acquisition means for acquiring the behavior history of the viewer of the first site in association with each viewer,
The action history acquisition means acquires an action history related to information on the first site via the second site of a viewer of the second site as an action history via the second site,
In the behavior history storage unit, a behavior history regarding information on the first site information presentation space of a plurality of viewers of the second site and a behavior history regarding information on the first site via the second site are provided to each viewer. Associated with it,
The recommendation rule creation unit
Recommendation rules specific to the second site based on the action history related to information on the first site information presentation space of a plurality of viewers on the second site and the action history related to information on the first site via the second site And a means to create
Recommendation rules specific to a specific viewer based on an action history related to information on the first site information presentation space of a specific viewer of the second site and an action history related to information on the first site via the second site And a means to create
It has.

In one aspect, the recommended information selection unit and the information presentation unit include:
If the viewer of the second site is the first visitor, present the information selected based on the recommendation rules specific to the second site in the first site information presentation space,
When the viewer of the second site is a repeater, the information selected based on the recommendation rule unique to the viewer is presented in the first site information presentation space.

In one aspect, the recommended information selection unit and the information presentation unit include:
When the viewer of the second site is the first visitor, when the viewer selects information on the first site information presentation space, the first information presentation space is based on the recommendation rules specific to the second site. The first site information presented in is selected and presented in real time,
When the viewer of the second site is a repeater, when the viewer selects information on the first site information presentation space, the information is presented in the first information presentation space based on a recommendation rule unique to the viewer. The first site information is selected and presented in real time.

In one aspect, the recommendation rule creation unit includes:
The creation of the recommendation rule specific to the second site is a means for acquiring a correlation coefficient between information based on the action history of a plurality of viewers of the second site;
The creation of the recommendation rule specific to the specific viewer is to obtain a correlation coefficient between information based on the behavior history of the specific viewer of the second site;
Is included.

In one aspect, the recommendation rule creation unit includes:
The creation of the recommendation rule specific to the second site is a means for obtaining a ranking of information that is easily selected on the second site based on the action history of a plurality of viewers on the second site;
The creation of the recommendation rule unique to the specific viewer includes a ranking of information that can be easily selected by the viewer based on the behavior history of the specific viewer on the second site,
The recommended information selection unit and the information presentation unit are:
When the visitor of the second site is the first visitor, when visiting the second site, the information selected based on the ranking of information that is easy to select on the second site is presented in the first site information presentation space,
When the viewer of the second site is a repeater, the information selected based on the ranking of information that can be easily selected by the viewer is presented in the first site information presentation space when visiting the second site.

  In one aspect, the first site is an EC site that is an advertiser, and the first site information presentation space of the second site is an advertising space for the first site. In one embodiment of the present invention, the second site is a partner site, and the information presented in the first site information presentation space is an affiliate advertisement.

In the present invention, the information on the first site information presentation space on the second site is optimized for each second site and each viewer of the second site, and for the first visitor of the second site. However, information that has already been optimized for the second site is presented, and information that is optimized more specifically for the viewer is presented to a plurality of viewers of the second site.

When the first site information presentation space on the second site is an advertising space, the distribution content is optimized according to the interest of the viewer, so that efficient advertisement distribution can be performed. When the advertising space is an affiliate advertisement, it is beneficial for the advertiser (operator of the first site) and the partner (operator of the second site).

1 is an overall view of a presentation information optimization system according to the present invention. It is a figure which shows embodiment (affiliate advertisement) using this invention. It is a figure explaining optimization of goods presented in an affiliate frame on a partner site. At the initial stage of distribution, products that are optimized by recommendation rules of the EC site are presented in the affiliate frame. By learning which of the presented products are actually clicked, products that are easily clicked on the partner site are selectively distributed (optimization for each partner site). It is a figure which shows the mechanism of the optimization of the advertisement shown by the affiliate frame on a partner site. The visitor's action history for the affiliate advertisement on the partner site is transmitted to the recommendation engine, the user's interest is estimated by the recommendation engine using the action history, and the optimum advertisement is presented in the affiliate frame on the partner site. It is a figure explaining the optimization for every partner site of the advertisement shown by the affiliate frame on a partner site. In site A, where there are many female visitors to the partner site, by accumulating the user's behavior history for the advertisement presented in the affiliate frame, an advertisement that is likely to be of interest to the users of site A is selected and presented. . Similarly, in Site B, where visitors to the partner site are more fond of cameras, it is possible to select advertisements that are likely to be of interest to users of Site B by accumulating the user's behavior history for advertisements presented in the affiliate window. And present it in the affiliate window. It is a figure explaining the optimization for every viewer of the advertisement shown by the affiliate frame on a partner site. When a visitor clicks on an advertisement, which advertisement is clicked is transmitted to the recommendation engine in association with the viewer, and an advertisement that is likely to be of interest to the individual viewer is extracted and presented in the affiliate frame. It is a figure explaining a recommendation rule. It is a figure which shows the transition of the banner (product advertisement presentation space) on a partner site (2nd site). It is a figure which shows other embodiment (movie ticket reservation and sales terminal) using this invention. It is a figure which shows other embodiment (Online game shop scenario selection) using this invention.

[A] Overall Structure of Presentation Information Optimization System FIG. 1 shows an overall view of a presentation information optimization system according to the present invention. The optimization system includes a first site that provides a plurality of information, a second site in which a first site information presentation space for presenting information on the first site is provided on a display screen, and a mediation server, The first site, the second site, and the mediation server are connected to each other via a computer network so that data can be transmitted and received. The mediation server provides a recommendation engine.

The intermediary server includes a behavior history storage unit that stores a viewer's behavior history in association with each viewer, a recommendation rule creation unit that creates a recommendation rule for selecting information recommended to the viewer using the behavior history, And a recommendation rule storage unit for storing the created recommendation rule.

In the behavior history storage unit, the behavior histories relating to information on the first site information presentation space of the viewer of the second site are acquired in association with each viewer, so that the first sites of the plurality of viewers of the second site An action history related to information on the information presentation space is stored in association with each viewer. The first site is provided with behavior history acquisition means for acquiring the behavior history of the viewer of the first site in association with each viewer, and the behavior history of the viewer of the first site is acquired in association with the viewer. The The action history acquisition unit further acquires an action history related to information on the first site via the second site of the viewer of the second site as an action history via the second site. In the action history storage unit, action history related to information on the first site information presentation space of a plurality of viewers of the second site and action history related to information on the first site via the second site are unique to the second site. It is stored as history data in association with each viewer. That is, the visitor history data, the first site history data, and the second site history data are stored in the action history storage unit of the mediation server.

The recommendation rule creating unit includes means for creating a recommendation rule unique to the first site based on the action history of the plurality of viewers on the first site, and the information presentation space of the plurality of viewers on the second site. A means for creating a recommendation rule specific to the second site based on an action history relating to the information of the second site, and a specific browsing based on the action history relating to information on the information presentation space of a specific viewer of the second site And a means for creating a recommendation rule specific to a person. In one aspect, the recommendation rule creating unit includes an action history related to information on the first site information presentation space of a plurality of viewers of the second site and an action history related to information on the first site via the second site. On the first site via the second site, the action history regarding the information on the first site information presentation space of the specific viewer of the second site and the means for creating a recommendation rule specific to the second site And a means for creating a recommendation rule specific to a specific viewer based on an action history related to the information. That is, the recommendation rule storage unit of the mediation server stores the viewer-specific rule data, the first site-specific rule data, and the second site-specific rule data.

The mediation server further includes a recommendation information selection unit that selects recommendation information to be presented to the viewer by appropriately selecting and using rule data stored in the recommendation rule storage unit, and a first site information presentation space of the second site. A presentation information storage unit for storing information relating to information on the first site for presenting information on the first site, and information selected by the recommended information selection unit from the presentation information storage unit to the information presentation space on the second site And an information presenting means for presenting. The recommendation information selection unit, the presentation information storage unit, and the information presentation unit are also used for presenting recommendation information to the viewer of the first site.

The optimization system of the presentation information which concerns on this invention is demonstrated based on the optimization system of the affiliate advertisement which is a suitable embodiment. As shown in FIG. 1A, the affiliate advertisement optimization system includes an EC site X, an intermediary server that provides a recommendation engine, and a partner site in which a space for selectively presenting product advertisements on the EC site X is set on the display screen. And. Product advertisement of EC site X is presented in the space on the partner site display screen via the mediation server, and the partner site viewer clicks the product advertisement on the space to link to EC site X and purchases the product. Etc., a reward is paid from the EC site X operator to the partner site operator. As a charging method, a conventional affiliate advertisement method can be used.

The EC site X, the partner site, and the mediation server are configured from one or a plurality of computers. The computer includes processing means as hardware, storage means, input means, output means or display means, a control program for operating the computer as software, and the like. EC site X, intermediary site, and partner site are equipped with transmission / reception means that enables mutual exchange of information via a computer network represented by the Internet, and can communicate with each other via a computer network represented by the Internet. It is connected to the. The EC site X and the partner site are typically websites, and are connected to user terminals via the Internet so as to be able to communicate with each other. The user terminal includes a computer, and includes processing means as hardware, storage means, input means, output means or display means, a control program for operating the computer as software, a browser for browsing, and the like.

The EC site X is typically a product sales site on the Internet, a so-called mail-order site, and a user terminal is connected to the EC site X via the Internet so as to be communicable. The storage unit of the EC site X stores a product information table related to products provided on the EC site X, and the product information table stores all of the information to be displayed on the web page. For example, information such as product name, product image, product description, and price is stored in the product information table. In response to the EC site X browsing request from the user terminal, a page related to the requested product is displayed on the display screen of the user terminal. A user who visits the EC site X browses a desired product page and purchases a product by inputting from an input means (such as a mouse) of the user terminal.

The EC site X is provided with an action history acquisition means for acquiring a visitor's action history (product browsing, product purchase, etc.), that is, an application for collecting the viewer's log data. B, C... Action history (product browsing, product purchase, etc.) is collected as an action log by the action history acquisition means. In the EC site X, the viewer's action log is transmitted from the EC site X to the mediation server via the computer network. Applications that collect log data of website visitors are known to those skilled in the art and are used in practice. For example, the period until a viewer accesses a website and leaves is handled in units of “sessions”, and browsing history and purchase history are acquired around “sessions”.

On the display screen of the partner site (website, e-mail magazine, etc.), a space for selectively presenting multiple product advertisements on the EC site X is set, and the multiple product advertisements are products on the display screen of the partner site. Present it in the advertisement presentation space. Each product advertisement on the product advertisement presentation space is directly or indirectly linked to a related product page of the EC site X. When a viewer of the partner site clicks on a product advertisement in the product advertisement presentation space of the partner site displayed on the display screen of the user terminal, a request to browse a web page related to the product on the EC site X is made, and the user terminal Transitions to the web page of EC site X. The browsing request for the web page of the EC site X may be made via (redirecting) an intermediary server or another server (for example, a server of an affiliated company). In addition, the EC site X may be installed in the mediation server. In this case, the flow is a partner site → mediation server → a product page of the EC site X installed in the mediation server. The specific configuration of the product advertisement presentation space on the display screen of the partner site is a matter that can be appropriately determined by those skilled in the art.

In response to the partner site browsing request input from the user terminal, the web page of the requested partner site is displayed on the display screen of the user terminal. When the user D visiting the partner site clicks on one of the product advertisements displayed in the product advertisement presentation space, a request to view a web page related to the product on the EC site X linked to the product is made, and the user The web page is displayed on the display screen of the terminal. As a result, the user D browses the EC site X with an incidental condition of “via partner site”. Even for the viewer D who flows in from the product advertisement presentation space of the partner site, log data when browsing / purchasing behavior is performed on the EC site X is collected by the behavior history acquisition means in association with the viewer D who has flowed in Then, it is transmitted to the mediation server via the computer network. The browsing history of the user D is collected by the mediation server, and is used to create a recommendation rule specific to the partner site and a recommendation rule specific to the user D.

The intermediary server receives an action history acquired at the EC site X, stores the action history in association with each viewer, and browses the EC site X using the action history stored in the action history storage unit. A recommendation rule creating unit (simultaneous browsing rule calculation between products in FIG. 1A) for obtaining a recommendation rule for selecting information recommended for a viewer, and a recommended information selection unit for selecting information recommended for a viewer using the recommendation rule (Affiliate optimization rule in FIG. 1A), a presentation information storage unit for storing product information of EC site X for presenting product information of EC site X in the product advertisement presentation space of EC site X of the partner site, and a recommendation A predetermined number of product advertisements are selected using the recommendation rule acquired by the rule creation unit, and the product information related to the selected product advertisement is selected. The by linking the EC site X from the presentation information storage unit, and a means for presenting the information presentation space on the partner site (advertisement display system in FIG. 1A), the.

As described above, the behavior history acquisition unit acquires the behavior history regarding the information on the EC site X via the partner site of the partner site viewer in association with the partner site and the viewer as the behavior history via the partner site. This action history is transmitted to the mediation server and stored in the action history storage unit. In addition, the action history related to the product advertisement on the product advertisement presentation space of the partner site viewer (clicking on the product advertisement) is transmitted directly to the mediation server for each click, and the partner site and the viewer are stored in the action history storage unit. Stored in association with. Therefore, in the action history storage unit of the mediation server, in addition to the action history of the viewer who directly accesses the EC site X, the action history and the partner site regarding the information on the information presentation space of the plurality of viewers of the partner site are stored. The action history regarding the information on the EC site X that has been passed through is stored in association with each viewer and each partner site. In the breakdown of the logs used when calculating the recommendation rules related to the product advertisement presentation space on the partner site, only the log for the EC site X in the initial state, but as the time passes, the log on the product advertisement presentation space on the partner site As a result, it will be optimized for partner sites. In the optimization, the log of the viewer who entered the EC site X via the partner site is also taken into account.

The presentation information storage unit of the mediation server stores the product information of the products handled on the EC site X, and the product information on the EC site X can be presented in the product advertisement presentation space on the display screen of the partner site. ing. The merchandise information of the EC site X in the mediation server is updated as appropriate, and corresponds to the latest merchandise provided by the EC site X. The presentation information storage unit stores the product information to be written when the product information related to the product page of the EC site X is written in order to present the product information in the product advertisement presentation space of the partner site. The type of information stored in the product information table of the mediation server differs depending on what product information is presented in the product advertisement presentation space on the display screen of the partner site. For example, if only the product name is described in the product presentation space, the product information may be a list of product names of all products on the EC site X. Alternatively, the product information table of the mediation server may store the same product information as the product information table of the EC site X. For example, when a viewer of the partner site clicks a certain product advertisement on the product advertisement presentation space, the related product detail page may be read from the mediation server. It is understood by those skilled in the art that what data is stored in the presentation information storage unit depends on the content data provided by the media or the site. For example, in the case of a news site, the title of the news and its contents are stored.

The recommendation rule creation unit of the mediation server creates a recommendation rule by calculating simultaneous browsing rules between information (products) using log data collected on the website. In the present embodiment, in one aspect, the recommendation rule creation unit includes a recommendation rule for a viewer of the EC site X, a recommendation rule for a viewer of the partner site, and a recommendation rule for a specific viewer of the partner site. Create a kind of recommendation rule. More specifically, the recommendation rule creation unit of the mediation server includes a means for creating a recommendation rule for the viewer of the EC site X based on the action history of the plurality of viewers of the EC site X, and a plurality of partner sites. A means for creating a recommendation rule specific to a partner site based on an action history related to information on the information presentation space of the viewer and an action history related to information on the EC site X via the partner site, and identification of the partner site Means for creating a recommendation rule specific to a specific viewer based on an action history related to information on the information presentation space of the viewer and an action history related to information on the EC site X via the partner site; I have. The recommendation rules will be described in detail later.

The recommended information selecting unit selects a predetermined number of product information of the EC site X presented in the partner site product advertisement presentation space based on the recommendation rule, and the information presenting means selects the product information selected by the recommended information selecting unit. Present in the partner site product advertisement presentation space. More specifically, the recommended information selection unit selects and updates a predetermined number of product advertisements on the EC site X to be presented in the product advertisement presentation space on the EC site X on the partner site based on a recommendation rule specific to the partner site. And a predetermined number of EC sites X to be presented in the product advertisement presentation space of EC site X on the partner site based on a means for optimizing the displayed product specific to the partner site and a recommendation rule specific to a specific viewer. Means for optimizing the product advertisement presented by selecting and updating the product advertisement specific to a specific viewer of the partner site. As described above, the product information presented in the product advertisement presentation space of the EC site X by the information presenting means is provided from the presentation information storage unit of the mediation server.

In this system, viewers are specified as necessary. There are several known methods for identifying viewers on the Internet, such as "How to have identification information on the terminal side and refer to and match on the next visit", "Allow viewers to log in, etc." “Method” can be exemplified. For the former, a unique ID such as a cookie for a PC and a terminal ID for a mobile can be used. For the latter, if the login is confirmed, the session can be specified based on the login information. By storing “from which partner site's product presentation space” the cookie is stored, the rule data of the partner site can be acquired even when the EC site X is accessed next time. Similarly, a specific viewer can be personalized. Cookies and unique ID information are linked to the log saved on the mediation server, and when the log that has fallen in a new session can be matched with that information, the log up to the previous time is displayed. Recommendation rules can be updated while taking over.

[B] Recommendation Rules In this embodiment, action histories of a large number of viewers on the EC site X are acquired, and recommendation rules are created from the acquired action histories. The browsing history of the browsing person of EC site X and the action history of the purchase log are transmitted to the mediation server. In this embodiment, the recommendation engine is provided in a mediation server independent of the EC site X. The intermediary server automatically acquires the “browsing log” and “purchasing log” of EC site X, and which information (product) and which information (product) are easy to be viewed simultaneously, which information (product) and which The recommendation information is automatically displayed by calculating a “recommendation rule” indicating whether information (product) is likely to be purchased (or downloaded or requested for materials) at the same time. Log data received from the EC site X, which is digital media managed by the advertiser, is stored in the mediation server in advance. Using the accumulated log data, recommendation rules between information (products) are calculated and stored.

It has already been done to recommend products based on the visitor's behavior history on the website. For example, if you browse or purchase a product on the website, "The person who saw this product is looking at the following product", "The person who bought this product also purchased the following product The product is recommended. It will be understood by those skilled in the art that various methods can be taken as to how to create a recommendation rule from a viewer's action history.

In creating recommendation rules, there are basic views such as browsing co-occurrence and purchasing co-occurrence. The browsing co-occurrence will be described with reference to FIG. A visitor A of a site browses products ○, products △, products □, a visitor B browses products ○, products △, products ★, and a visitor C views products △, products □, products ▼. And It is assumed that the new visitor D browses the product Δ when these browsing histories exist. It is the product ○ and the product □ that have been confirmed to co-occur with the product Δ in the browsing history, and therefore the product ○ and the product □ are recommended. In this way, using the browsing log acquired and accumulated within a predetermined period, association analysis based on browsing history of which information (product) and which information (product) were browsed simultaneously (one session) The correlation coefficient between products (information) is acquired using. The same can be said about purchasing co-occurrence.

Although the creation of recommendation rules using association analysis was simplified, the correlation coefficient between each product was calculated using the ratio of the number of times each product was viewed to the total number of times viewed, the frequency of co-occurrence between each product, etc. One skilled in the art will appreciate that can be calculated. Further, the correlation coefficient may be calculated only from the browsing history, or the correlation coefficient may be calculated using both the browsing history and the purchase history. It will be understood by those skilled in the art that, in the latter case, when the correlation coefficient is calculated, appropriate weighting can be performed between the browsing history and the purchase history. In addition, the recommendation rule using association analysis expresses the relationship of “what information is easy to view or review at the same time”, and not only the one-to-one relationship but also the information group (product group). It is also possible to represent the relationship between information groups (product groups) that are easily viewed or considered at the same time as keys.

Creation of a recommendation rule is not limited to acquisition of a correlation coefficient using association analysis. The creation of the recommendation rule includes obtaining a ranking of information that can be easily selected on the site based on the browsing history of the viewer on a certain site. For example, when the initial presentation information (product) when a visitor visits a site is selected and presented based on the ranking (first recommendation rule) and the viewer clicks on the information (product) Information (product) may be selected and presented using the correlation coefficient (second recommendation rule).

Information used to create recommendation rules includes action data (clicks, orders, etc.) generated by the target person receiving the recommendation, action history data of the target person (past browsing history and purchase history), There are reporting data (questionnaire data and reporting data), merchandise information data provided on the EC site X, action history data as a whole of all viewers (including target users), and the like. Depending on what information is presented based on what way of thinking, the creation of recommendation rules may differ. In the present embodiment, a recommendation rule is created by mainly using a viewer's action history, but a recommendation rule may be created taking into account information other than the action history. For example, when the viewer's attribute is known, the recommendation rule may be calculated using the viewer's attribute as one of the parameters. It is understood that recommendation rules can be appropriately designed by those skilled in the art according to the characteristics of the partner site, the characteristics of the products handled at the EC site X, and the like. It is understood by those skilled in the art that a recommendation rule can be created by a computer by digitizing each parameter and expressing the recommendation rule in the form of a mathematical expression.

Several types of recommendation rules will be described. There are Attribute (Mono attribute base), Consumers (Human base), and Knowledge (Knowledge base). The thing attribute base is a method of presenting a product based on the idea that “a person evaluates a thing comprehensively by evaluating each attribute of the thing”. The person base is, "People with similar characteristics and those with similar past purchase histories prefer the same thing.""Products and information that many people considered and purchased at the same time are subject to the target. This is a technique for deciding which products to present based on the concept of “very valuable products and information”. The knowledge base is valuable to the target person based on the knowledge and experience of third parties (not limited to specialists and shop assistants) and general common knowledge (knowledge), regardless of whether or not it is related to the taste of the target person. It is a technique for presenting possible products. In one embodiment, the recommendation method can be classified into 12 types as shown in Table 1. By this classification, it is possible to examine an optimal recommendation method according to the merchandise to be handled and the data acquisition status. However, the classification here is merely a categorization of the basic concept, and in fact, a hybrid recommendation method combining a plurality of types can be designed. In Table 1, ACK indicates an initial of three types of recommendation logic: Attribute (mono attribute base), Consumers (human base), and Knowledge (knowledge base).

A-1. This is a method of determining products to be presented based on the specifications and characteristics of specifications ranking type products. For example, “present products in order of price”, “present products in order of price”, and the like. It can be used as an initial product selection in the product presentation space of the partner site. For example, if it is known in advance that the viewer of the partner site prefers low-priced products, a predetermined number of products are presented in order of price.

A-2. Product relevance evaluation type This is a method of presenting products that are closely related to the product to a person who checks or purchases the product. The depth of association is defined by the similarity and mutual complementarity of products. For example, presenting “a product whose performance is close to that of the checked product” or “a consumable necessary for the purchased product” may be presented. This can be used when the degree of similarity and mutual complementarity of products can be defined in the products provided on the EC site X.

A-3. This is a method to extract the features of products that the target user likes from the product list that the deep knowledge type target has checked or purchased in the past, and extract and present the products having the same kind of features from the product database. It is assumed that the characteristics of the product can be converted into data and that there is sufficient purchase history data of the target individual.

A-4.Needs for the needs input type Target audiences are listened to in advance, and products with the ideal characteristics required by the target are defined. This is a technique in which the ideal product is matched with the characteristic database of the recommended target product and presented in order from the product that seems to be closest to the ideal product of the target person. Because it directly asks the needs of the target audience, high-precision recommendations can be made, but know-how is needed to set up the question items and listening methods that do not put stress on the target audience, and at the same time, a product database can be built It is a premise.

C-1. It is a simple method of presenting popular products of popularity ranking type. This includes introductions of popular products, top-ranked products, and information that has been viewed frequently. It can be used as an initial product selection in the product presentation space of the partner site. For example, a predetermined number of products with a large number of browsing by the viewers of the partner site are presented in the order of browsing count.

C-2. Action association type Using data that a target person checks or purchases a product and data that other than the target person checks or purchases, perform association analysis (correlation analysis) on the co-occurrence between products based on the purchase pattern. It is a technique to derive and present using. For example, a presentation “a person who bought this product also bought such a product” is included in this category. However, in this type, any person who checks or purchases the same product will be presented with the same product, so all personal past purchase histories are learned and personalized collaborative filtering, etc. It is different from the history association type. Also, in the case of durable consumer goods, people who are considering this refrigerator are interested in this refrigerator using the simultaneous examination data compared at the time of purchase consideration.

C-3. History association type (collaborative filtering)
Using data that a target person has checked or purchased a product and data that has been checked or purchased by someone other than the target person, an association analysis (correlation analysis) of similarity between people or co-occurrence between products from the purchase pattern It is a technique for presenting personalized products by deriving and using personal behavior history. It is characterized by being personalized using the history data of the target individual. A typical analysis algorithm is “collaborative filtering”. Collaborative filtering is based on the hypothesis that people who have similar preferences and purchase histories like the same thing, and extract people with behavior histories similar to the target user's behavior history. This is a method of presenting products that the people are interested in or have bought and that the target person has not yet met. It is characterized by defining similar people and presenting what they have purchased but not the target. Even if there is no product database, it can be executed if there is only purchase history data other than the target person and the target person.

C-4. Questionnaire-based type Based on the subject attributes (gender, age, occupation, interests, etc.) that the subject has previously declared, people who have the same or similar attributes as the subject or those who belong to the same cluster It is a recommendation that presents products that have been purchased a lot. It is necessary to separately analyze data regarding which products consumers of each attribute are interested in or which products are purchased. For sites that require membership registration, general target person attributes can often be obtained. In this case, the tendency of purchase behavior for each attribute can be extracted from analysis of past purchase data. Even when past purchase data does not exist, it is possible to extract a tendency for each attribute by separately conducting a questionnaire survey or the like, and determine a product to be presented for each attribute therefrom.

K-1. This is a method that uniformly presents universal recommendations that knowledge-based specialists, employees, and salespeople think. “Our recommended products”, “Today's recommended products”, “Critic critic Mr. XX's recommended products” and so on apply. A certain effect can be expected if the reliability and validity of the recommendation grounds can be secured.

K-2. Knowledge base reaction type third party (not necessarily an expert / salesperson) is a method of presenting products that are considered preferable to those who have checked or purchased a product at their own discretion. is there. A certain effect can be expected if the reliability and validity of the recommendation grounds can be secured.

K-3. As a result of observing the past behavior history and purchase history of the knowledge base observation type target person, it is recommended by a certain standard by the independent judgment of a third party (not necessarily an expert / salesperson) This is a technique for presenting products and information that should be used. For example, we recommend vegetables to people who have only purchased meat, considering the nutritional balance.

K-4. According to the attributes (gender / age / profession / interest), needs, and concerns reported by the knowledge base diagnosis type subject, it is considered to be optimal based on certain criteria based on knowledge from experts, employees, salespersons This is a technique for presenting products that are available.

For example, a recommendation rule of A-1 or C-1 is used for a visitor who first visited the site after the partner site has been optimized several times. Alternatively, the initial information presented first may be arbitrarily selected and presented by the operator of the second site, for example. If you click once in that state, the recommendation rules A-2 and C-2 are used. Furthermore, if personalized data is accumulated, such as proceeding to purchasing behavior, it is possible to use recommendation rules of A-3 and C-3. Recommending a product using a recommendation rule has already been performed, and in the present embodiment, the content of the recommendation rule itself is not limited. What kind of recommendation rule is created and used is a matter that can be appropriately designed by those skilled in the art. It is understood that the recommendation rule update frequency can be appropriately set by those skilled in the art. As described below, the present embodiment is characterized in that two optimizations are performed. In one aspect, a different recommendation rule is used in each optimization.

[C] Optimization of presented advertisement In the embodiment of the present invention, a plurality of advertisement target contents handled by the advertiser are displayed on the advertisement medium, but the advertisement target contents are viewed by the following two methods. Optimize for

[C-1] In an optimized advertising medium (partner site) for each advertising medium, every time a viewer selects advertising target content, the action log is transmitted to the mediation server. This makes it possible to grasp what kind of advertising target content has been selected on the partner site. Each partner site is regarded as a single user, and affiliate advertisement click logs are accumulated, and products that are most easily clicked are selectively displayed based on actual data for each partner site. The intermediary server determines the content that can be easily selected on the partner site by referring to the recommendation rule using the past advertisement target content selected on the partner site as a key, and uses the advertisement target content data stored in advance. By switching and presenting advertisements to be posted on the partner site, it can be optimized specifically for the partner site (FIG. 4). Therefore, after a certain viewer selects and optimizes the advertising target content on the partner site, the optimized advertising target content (product) is presented to the first viewer who visits thereafter. It becomes a state.

[C-2] Optimization for each viewer (personalization)
When the viewer selects the advertisement target content in the partner site, the advertisement target content (product) displayed for the viewer is changed and optimized in real time (FIG. 5). For example, it is assumed that a certain viewer selects a certain advertisement target content (content α) on the partner site. The recommendation engine of the mediation server refers to the recommendation rule in real time, and displays content that is likely to be interesting simultaneously with the content α on the partner site. In this way, while the viewer is looking at the partner site, the advertising target content displayed every time the viewer selects the advertising target content (product) is optimized.

With these two types of optimization, the recommendation engine can display advertisements that are likely to be of interest to viewers. That is, in this embodiment, the product advertisement on the product advertisement presentation space of the EC site X on the partner site is optimized for each partner site and each viewer of the partner site. Also, information that has already been optimized for the partner site is presented, and information that is more specifically optimized for the viewer is presented to the viewer of the partner site multiple times. Therefore, since the product advertisement is optimized according to the interest of the viewer, efficient advertisement distribution can be performed. When the advertising space is an affiliate advertisement, it is beneficial for providers of both the partner site and the EC site X.

[D] Flow of affiliate advertisement optimization One aspect of the overall flow of affiliate advertisement optimization will be described with reference to FIG. 1A. The flow described below is only an example and does not limit the present invention. Moreover, it will be understood by those skilled in the art that the steps are not necessarily executed in the order described, and the described steps can be interchanged.

An application for collecting action history data is introduced to EC site X. When the viewer A who has accessed the EC site X browses and purchases a product, the behavior history is collected by the behavior history data collection application, transmitted to the mediation server, and stored in the behavior history storage unit. Similarly, the behavior histories of a plurality of viewers B, C,... Are collected by the behavior history data collection application, transmitted to the mediation server, and accumulated in the behavior history storage unit. Further, the product data of the EC site X is also transmitted to the mediation server and stored in the presentation information storage unit.

Based on the action history collected from EC site X only, the recommendation engine of the mediation server uses the recommendation engine of the intermediary server with the guideline that “the person who is interested in the same product is also interested in the other product that one person saw”. The recommendation rule (simultaneous browsing rule between products) in the site X is created, and when a certain product is browsed, preparation is made so that another product information can be recommended.

A space for presenting a product advertisement for EC site X is set on the display screen of the partner site. The product advertisement is selected by the recommended information selection unit, and is presented in a space where the information presentation unit presents the product advertisement using the product data in the presentation information storage unit.

The product advertisement presented in the product advertisement presentation space of the partner site is selected and presented based on the ranking data of the EC site X or similar rules in the initial state immediately after setting.

When the viewer D of the partner site selects (clicks) the product advertisement, the history is transmitted to the mediation server and accumulated in association with the viewer D in the action history storage unit. In the initial state immediately after the setting, the recommendation information selection unit selects the recommended product advertisement using the recommendation rule based on the log data of the EC site X.

When the viewer D flows into the EC site X from the partner site via the product advertisement presentation space, the log acts by acting in the EC site X with the incidental condition of “flowing in from the partner site” It is acquired by the history collection application, transmitted to the mediation server, and accumulated in association with the viewer D in the action history storage unit. When calculating the recommendation rule, this log is added to the log on the product advertisement presentation space of the partner site, and optimization is performed for each partner site and each specific viewer. It should be noted that the EC site X collects all action histories, regardless of whether or not incidental conditions are attached, and uses them for calculating recommendation rules for the EC site X.

Using the action history accumulated in the action history storage unit, recommendation rules (recommendation rules unique to the partner site and recommendation rules specific to a specific viewer) are created by the recommendation rule creation unit on the partner site (product advertisement presentation space) And reflected in the selection of the product advertisement in the product advertisement presentation space on the partner site.

When one of the product advertisements presented by the partner site viewer D is selected and clicked, the product information to be displayed from the click log is inquired to the recommended information selection unit of the broker site, and the click log and the viewer D specific The product advertisement to be presented to the viewer D is selected using the recommendation rule, and the selected product advertisement is displayed in real time by the information presenting means (personalization). The recommendation rule unique to the viewer D personalized by the viewer D does not affect the recommendation rule personalized by the other viewers.

After that, each time another visitor visits the partner site and clicks on the product displayed on the partner site display screen, the recommendation rule specific to the partner site and the personalized viewer-specific Recommendation rules are created and updated. In addition, the latest recommendation rule of EC site X may be periodically added to the recommendation rule for each partner site and each viewer.

Based on FIG. 7, the transition of the banner (product advertisement presentation space) on the partner site (second site) will be described. As shown in the left diagram of FIG. 7, in the product advertisement presentation space on the display screen of the partner site, as an initial state, the product A is presented on the upper side as a large region, and the products B to G are on the lower side as small regions. Presented. In the initial state, when the viewer of the partner site selects the product B, the product B is presented in the upper large area. When the product B is selected, the product presented on the lower side is updated to the recommended products E, G, A, D, C, and H according to the recommendation rule with the product B as an input.

When the viewer of the partner site clicks the purchase button or link button of the product B (for example, formed in the lower area of the banner), the link is made to the product B page of the EC site X, and the display screen of the viewer's terminal Transitions to the product B page of C site X. If the viewer requests the page of the product Y linked to the page of the product B without purchasing the product B, and purchases the product Y, the browsing history of the products B and Y and the purchase history of the product Y are the mediation server. And is reflected in the recommendation rule. It will be understood by those skilled in the art that the layout of the product presented in the product advertisement presentation space is not limited to the aspect of FIG.

[E] Other Embodiments The method and system for optimizing presentation information is not limited to the above embodiment. For example, on-demand broadcast program introduction on TV, ticket reservation system for trains, airplanes, various performances, library OPAC, operator PC for call center operations, shopping screens for online games, scenario selection, etc. Is possible.

For example, a business model (movie ticket reservation / sales terminal) as shown in FIG. 8 is possible. In addition to movie theaters, ticket reservation terminals for movies are installed in ticket information stores such as Ticket Pia and convenience stores. By displaying the recommendation information obtained on the conventional PC site or member database here, it is possible to promote the improvement of customers and conversion.

Further, a business model (online game shop scenario selection) as shown in FIG. 9 is possible. Online games, by their nature, have a database that can completely identify individuals. Moreover, even if many people participate, only one terminal is accessed per person. Using this, it is possible to optimize or personalize from the logs of a large number of users, thereby promoting the purchase of goods and the activation of game play.

Claims (12)

A first site information presentation space for presenting information of the first site providing a plurality of information is provided on the display screen of the second site, selecting information from the plurality of information of the first site, and selecting the selected information Presenting in the first site information presentation space on the second site;
The action history regarding the information on the first site information presentation space of a plurality of viewers of the second site is acquired, and a recommendation rule specific to the second site is created using the action history, and the recommendation rule specific to the second site Selecting and updating the first site information presented in the first site information presentation space on the second site based on
An action history related to information on the first site information presentation space of a specific viewer of the second site is acquired, and a recommendation rule specific to the specific viewer is created using this action history, and a specific recommendation specific to the viewer Selecting and updating the first site information presented in the first site information presentation space on the second site based on the rules;
A method for optimizing presentation information using a computer network equipped with a computer. The first site information is linked to the first site and presented in the first site information presentation space on the second site,
In addition to the action history related to information on the first site information presentation space of a plurality of viewers of the second site, the action history related to information on the first site via the second site is acquired, and these action histories are obtained. To create recommendation rules specific to the second site,
In addition to the action history related to the information on the first site information presentation space of the specific viewer of the second site, the action history related to the information on the first site via the second site is acquired, and these action histories are To create specific viewer-specific recommendation rules,
The method for optimizing presentation information according to claim 1. If the viewer of the second site is the first visitor, present the information selected based on the recommendation rules specific to the second site in the first site information presentation space,
When the viewer of the second site is a repeater, the information selected based on the recommendation rule unique to the viewer is presented in the first site information presentation space.
The presentation information optimizing system according to claim 1, further comprising: When the viewer of the second site is the first visitor, when the viewer selects information on the first site information presentation space, the second site visitor changes to the first information presentation space based on the recommendation rule unique to the second site. Select and present the first site information to be presented in real time,
When the viewer of the second site is a repeater, when the viewer selects information on the first site information presentation space, it is presented in the first information presentation space based on a recommendation rule unique to the viewer. The first site information is selected and presented in real time.
The method for optimizing presentation information according to claim 1. The creation of the recommendation rule specific to the second site includes obtaining a correlation coefficient between information based on the action history of a plurality of viewers of the second site,
The creation of the recommendation rule specific to the specific viewer includes acquisition of a correlation coefficient between information based on the behavior history of the specific viewer of the second site. Optimization method. The creation of the recommendation rule specific to the second site includes obtaining a ranking of information that is easily selected on the second site based on the action history of a plurality of viewers on the second site,
Creation of the recommendation rule specific to the specific viewer includes obtaining a ranking of information that can be easily selected by the viewer based on the behavior history of the specific viewer of the second site,
When the visitor of the second site is the first visitor, at the time of visiting the second site, the information selected based on the ranking of information that can be easily selected on the second site is presented in the first site information presentation space,
When the viewer of the second site is a repeater, when visiting the second site, the information selected based on the ranking of information that can be easily selected by the viewer is presented in the first site information presentation space.
The method for optimizing presentation information according to claim 1. A first site providing a plurality of information;
A second site provided on the display screen with a first site information presentation space for presenting information of the first site;
An intermediary server;
The first site, the second site, and the mediation server are connected to each other via a computer network so that data can be transmitted and received.
The intermediary server includes a behavior history storage unit that stores a viewer's behavior history in association with each viewer, a recommendation rule creation unit that creates a recommendation rule for selecting information recommended to the viewer using the behavior history, A recommendation rule storage unit for storing the created recommendation rule, a recommendation information selection unit for selecting recommendation information using the recommendation rule, and the information on the first site in the first site information presentation space of the second site A presentation information storage unit for storing information related to the information on the first site, information presentation means for presenting information selected by the recommended information selection unit from the presentation information storage unit to the information presentation space on the second site,
With
The behavior history storage unit obtains behavior histories related to information on the first site information presentation space of the viewer of the second site in association with each viewer, so that the first of a plurality of viewers of the second site can be obtained. An action history related to information on the site information presentation space is stored in association with each viewer.
The recommendation rule creation unit
Means for creating a recommendation rule specific to the second site based on an action history related to information on the information presentation space of a plurality of viewers of the second site;
Means for creating a recommendation rule specific to a specific viewer based on an action history related to information on the information presentation space of the specific viewer of the second site;
With
The recommended information selection unit and the information presentation unit are:
Select and update the first site information presented in the first site information presentation space on the second site based on the recommendation rules specific to the second site,
Selecting and updating the first site information presented in the first site information presentation space on the second site based on a specific viewer-specific recommendation rule;
Presentation information optimization system. The first site is provided with behavior history acquisition means for acquiring the behavior history of the viewer of the first site in association with each viewer,
The action history acquisition means acquires an action history related to information on the first site via the second site of a viewer of the second site as an action history via the second site,
In the behavior history storage unit, a behavior history regarding information on the first site information presentation space of a plurality of viewers of the second site and a behavior history regarding information on the first site via the second site are provided to each viewer. Associated with it,
The recommendation rule creation unit
Recommendation rules specific to the second site based on the action history related to information on the first site information presentation space of a plurality of viewers on the second site and the action history related to information on the first site via the second site And a means to create
Recommendation rules specific to a specific viewer based on an action history related to information on the first site information presentation space of a specific viewer of the second site and an action history related to information on the first site via the second site And a means to create
With
The system for optimizing presentation information according to claim 7. The recommended information selection unit and the information presentation unit are:
If the viewer of the second site is the first visitor, present the information selected based on the recommendation rules specific to the second site in the first site information presentation space,
When the viewer of the second site is a repeater, the information selected based on the recommendation rule unique to the viewer is presented in the first site information presentation space.
The system for optimizing presentation information according to claim 7. The recommended information selection unit and the information presentation unit are:
When the viewer of the second site is the first visitor, when the viewer selects information on the first site information presentation space, the first information presentation space is based on the recommendation rules specific to the second site. The first site information presented in is selected and presented in real time,
When the viewer of the second site is a repeater, when the viewer selects information on the first site information presentation space, the information is presented in the first information presentation space based on a recommendation rule unique to the viewer. The first site information to be selected and presented in real time,
The presentation information optimization system according to claim 7. The recommendation rule creation unit
The creation of the recommendation rule specific to the second site is a means for acquiring a correlation coefficient between information based on the action history of a plurality of viewers of the second site;
The creation of the recommendation rule specific to the specific viewer is to obtain a correlation coefficient between information based on the behavior history of the specific viewer of the second site;
The optimization system of the presentation information in any one of Claims 7 thru | or 10 containing these. The recommendation rule creation unit
The creation of the recommendation rule specific to the second site is a means for obtaining a ranking of information that is easily selected on the second site based on the action history of a plurality of viewers on the second site;
The creation of the recommendation rule unique to the specific viewer includes a ranking of information that can be easily selected by the viewer based on the behavior history of the specific viewer on the second site,
The recommended information selection unit and the information presentation unit are:
When the visitor of the second site is the first visitor, when visiting the second site, the information selected based on the ranking of information that is easy to select on the second site is presented in the first site information presentation space,
When the viewer of the second site is a repeater, when visiting the second site, the information selected based on the ranking of information that can be easily selected by the viewer is presented in the first site information presentation space.
The system for optimizing presentation information according to claim 7.