JP2009245279A - Method and program for distributing advertisement - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To increase advertisement profits by optimizing the frequencies at which a plurality of advertisements assigned to a particular site are distributed. <P>SOLUTION: An advertisement distribution server 20 has stored therein an advertisement distribution program 21, an advertisement registration database 22, and an advertisement tallying database 23. The advertisement distribution program 21 includes an advertisement tallying module 21a for tallying the number of times that the plurality of advertisements distributed to the particular site are displayed, sales, etc., and recording the results in the advertisement tallying database 23, an optimization weight calculation module 21b for calculating an optimization weight that gets higher the greater the sales of an advertisement are, an initialization weight calculation module 21c for calculating an initialization weight that gets higher the newer an advertisement is, and an advertisement distribution module 21d which distributes, in a predetermined ratio, advertisements whose frequencies of distribution are determined by the optimization weight and advertisements whose frequencies of distribution are determined by the initialization weight. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a method for distributing advertisements to a transmission source for public transmission, and a program for causing a computer to distribute advertisements.

  With the spread of the Internet, advertisements posted on sites on the Internet have become an important element of product promotion. The advertiser requests the advertisement distribution company to place an advertisement, and the advertisement distribution company distributes the requested advertisement to a specific site. Each site has its own unique tendency depending on the content, so the audience of users who browse is also different. Therefore, if an advertisement that stimulates the purchasing interest of the audience is viewed, the rate at which the advertisement banner is clicked The higher the price, the more the advertiser pays to the advertising distribution company. For this reason, matching between a site and an advertisement is important.

  Conventionally, the person in charge of the advertisement distribution company determines the matching between the site and the advertisement as described above and, for example, selects to distribute an advertisement for sports equipment to a sports site. Then, when a plurality of advertisements are distributed in rotation in one advertisement space, the computer advertisement distribution system randomly distributes the plurality of advertisements so that the frequency of distribution is equal.

  On the other hand, a technique for selecting an advertisement to be distributed by paying attention to an attribute of a user browsing the site has been developed. For example, in Patent Literature 1, an advertisement distribution system displays advertisements in a field in which the user is interested on a site browsed by the user by selecting a field in which the user is interested in and registering it in the database in advance. A method is disclosed.

  Further, Patent Document 2 stores a sample user's access log, analyzes the sample user's access log that is close to the conditions of the target user of the advertisement, and obtains a site, period, and budget to which the advertisement should be distributed. It is disclosed.

JP 2002-049838 A JP 2001-306934 A

  However, simply distributing a plurality of advertisements assigned to a specific site in a random manner so that the frequency is the same as in the past, an advertisement with a large number of clicks and a large amount of advertisement rewards can also be paid with a small number of clicks. Since a small number of advertisements are distributed at the same frequency, there is a problem that it is not efficient for the advertisement distribution company.

  Further, in the systems disclosed in Patent Documents 1 and 2, the preferences of specific users must be registered in advance or the preferences of sample users must be stored in advance, and these user data are distributed. It is not an evaluation of the advertisement itself, but it is indirect, so it does not necessarily lead to an increase in advertising rewards.

  The present invention has been made in view of the above-described problems, and its problem is to optimize the distribution frequency of a plurality of advertisements allocated to a specific public transmission source, thereby reducing the advertisement reward. An object is to provide an advertisement distribution method and an advertisement distribution program that can be increased.

  In order to solve the above-described problem, an advertisement distribution method according to claim 1 of the present invention is a method of distributing advertisements to a transmission source of public transmission, and totals sales of a plurality of advertisements distributed to a specific transmission source. Calculate the percentage of total sales for each ad, determine the optimization weight for each ad based on the calculated percentage, determine the frequency of delivery for each ad by weighting the optimization weight, Each advertisement is distributed to the transmission source based on the determined frequency. Public transmission is a concept stipulated in Article 2, Paragraph 1, Item 7-2 of the Copyright Act. In “transmission of wireless communication or wired telecommunications for the purpose of direct reception by the public” The concept includes distribution via a computer network such as wired or wireless broadcasting or the Internet. Public transmission sources include, for example, websites on the Internet.

  Moreover, the advertisement delivery method according to claim 2 of the present invention is based on claim 1 and obtains the number of times each advertisement is displayed in the transmission source, and calculates the reciprocal of the ratio of the total number of advertisements distributed to the transmission source. Calculate the initial weight of each advertisement based on the calculated reciprocal, and distribute the advertisement whose distribution frequency is determined by the optimization weight and the advertisement whose distribution frequency is determined by the initial weight at a predetermined ratio It was made to do. For example, the ratio is 80% for the optimization weight and about 20% for the initial weight.

  An advertisement distribution program according to claim 3 of the present invention is a program for distributing advertisements to a transmission source of public transmission, an advertisement totaling procedure for totaling sales of a plurality of advertisements distributed to a specific transmission source in a computer, Calculate the proportion of each ad in the total sales, calculate the optimization weight for each ad based on the calculated proportion, and weight the ad weight to determine the frequency of each ad distribution. And an advertisement distribution procedure for distributing each advertisement to a transmission source based on the determined frequency.

  In addition, the advertisement distribution program according to claim 4 obtains the display frequency of each advertisement in the transmission source on the premise of claim 3, calculates the reciprocal of the ratio of the display frequency of all advertisements distributed to the transmission source, It further includes an initial weight calculation procedure for obtaining an initial weight for each advertisement based on the calculated reciprocal. In the advertisement distribution procedure, the distribution frequency is determined by the advertisement whose distribution frequency is determined by the optimization weight and the initial weight. It is characterized in that it is distributed at a predetermined rate.

  According to the advertisement distribution method and program of the present invention, since the appearance frequency can be determined based on the actual sales of advertisements, it is possible to efficiently distribute highly profitable advertisements and increase advertisement rewards.

  Also, if the distribution frequency is determined using the initial weight in addition to the optimization weight, new advertisements can be distributed at a certain rate. That is, if the distribution frequency is determined only by the optimization weight, it becomes difficult to distribute a new advertisement without a track record, but such a situation can be avoided by using the initial weight together. Since the initial weight becomes smaller as the number of display times increases, the distribution frequency by the initial weight becomes smaller as time passes. However, since the results are accumulated as the number of times of display increases, the optimization weight increases for a highly profitable advertisement, and the distribution frequency does not decrease.

Embodiments of an advertisement distribution method and an advertisement distribution program according to the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of a network system including an advertisement distribution server for realizing the advertisement distribution method of the embodiment. This system includes, as an example of a transmission source for public transmission, a plurality of sites deployed on the Internet N (generally indicated by reference numeral 10), a content server 11 that provides content to these sites, and these sites. It is comprised from the advertisement delivery server 20 which delivers an advertisement. The distributed advertisement is displayed on the web page of the site as an advertisement banner. An advertiser system 30 that orders advertisements is also connected to the Internet N.

  The user can access these sites 10 by using the user terminal 40 and browse the pages, and can link to the advertiser's site by clicking the advertisement banner displayed in the page. In this example, it is assumed that a mobile phone is used as the user terminal 40, and therefore the advertisement distribution site is also a mobile site.

  The advertisement distribution server 20 is a general-purpose computer and includes a CPU, a memory, a hard disk, an input device such as a keyboard and a mouse, and a monitor (not shown). An advertisement distribution program 21 is stored in the hard disk, and an advertisement is distributed to a web page of a site provided by the content server 11 by executing this program. In addition, the hard disk of the advertisement distribution server 20 stores an advertisement registration database (DB) 22 in which information of advertisements to be distributed is registered, and an advertisement totalization DB 23 that records the display frequency and sales of each distributed advertisement. ing.

  As shown in Table 1 below, the advertisement registration DB 22 has one record for each advertisement requested for distribution, and each record includes an advertisement identification number (ID), an advertiser, an advertisement content, and a unit price. Contains fields.

  As shown in Table 2 below, the advertisement tabulation DB 23 has a plurality of records for each advertisement by day of the week and by time. Each record has an advertisement ID, a site where the advertisement is distributed, a distribution day of the week, and a distribution time. Includes fields for bands, impressions, and sales. The time zone is set by dividing the day into five time zones: morning, noon, afternoon, night, and midnight. For example, from 6:00 to 10:59 in the morning, from 11:00 to 13:59 in the morning, from 14:00 to 17:59 in the afternoon, from 18:00 to 22:59 in the evening, from 23:00 to 5: 59.

  The advertisement distribution program 21 calculates an advertisement totaling module 21a that counts the number of times of display and sales of a plurality of advertisements distributed to a specific site and records them in the advertisement totalization DB 23, and an optimization weight that increases as the sales increase. An optimization weight calculation module 21b, an initial weight calculation module 21c for obtaining an initial weight that is higher for new advertisements, an advertisement whose distribution frequency is determined by an optimization weight, and an advertisement whose distribution frequency is determined by an initial weight And an advertisement distribution module 21d that distributes. The distribution ratio is, for example, 80% based on the optimization weight and 20% based on the initial weight.

Next, a method for obtaining the optimization weight and the initial weight will be described. The optimization weight is determined based on the calculated ratio by calculating the ratio of the sales of each advertisement to the total sales based on the information recorded in the advertisement tabulation DB 23. The sales of each advertisement is obtained as the value of the advertisement by the following calculation formula.
Value of advertisement = (M (site sales on the day) + N (total sales on the day) + (past sales)) / total number of impressions. The total sales are the sales on the day of all the sites where the advertisement is distributed, and the past sales are the total of the site sales and the total sales up to the previous day. M and N are coefficients for weighting and are fixed values. The value obtained by dividing the total of these three sales by the total number of display times is the value of the advertisement.

  For example, when the values of advertisements A to E distributed by rotation in a certain advertising space are obtained as shown in Table 3 below, the percentages and thousandths of each value with respect to these totals are obtained, and the thousandths are optimized. Used as a weight.

On the other hand, the initial weight is calculated based on the calculated reciprocal number by calculating the reciprocal of the ratio of the total number of advertisements distributed to a specific site from the display count recorded in the advertisement aggregation DB 23. Specifically, it is obtained by the following calculation formula.
Initial weight = (1− (total number of impressions of each advertisement / total number of impressions of all advertisements)) × 100

  For example, when the display frequency of advertisements A to E distributed by rotation in a certain advertisement space is obtained as shown in Table 4 below, the initial weight of each advertisement is calculated by the above calculation formula. The initial weight of each advertisement is calculated by the above formula. The distribution frequency based on the initial weight is determined by a percentage obtained by dividing the initial weight of each advertisement by the total initial weight.

  Next, the processing flow of each server based on the above advertisement distribution method will be described based on the flowcharts shown in FIGS. FIG. 2 shows the entire flow from the access by the user terminal 40 to the delivery of the advertisement to the site browsed by the user, and FIG. 3 shows the process of determining the delivery advertisement in the advertisement delivery server 20 shown in step S007 of FIG. Show details.

  As shown in FIG. 2, when the process starts and the user terminal 40 accesses a predetermined page of a specific site 10 (S001), the content server 11 issues a request for acquiring an advertisement to be displayed on the page. It is sent to the distribution server 20 (S002).

The advertisement distribution server 20 receives the request from the content server 11 (S003), and determines whether there is a record of advertisement distribution to the designated site 10 (S004). Here, for example, it is determined that there is no record if the record of advertisement distribution is less than 50, and that there is a record if it is 50 or more. If there is no record (S004, No), neither the optimization weight nor the initial weight takes a significant value, so that the advertisement is randomly extracted from the advertisement registration DB 22 so that the distribution frequency is equal to the conventional one (S005). Then, it is transmitted to the content server 11 (S006). If there is a track record (S004,
In (Yes), an advertisement is selected based on the weight and taken out from the advertisement registration DB 22 (S007),
It transmits toward the content server 11 (S006).

  Here, details of the advertisement extraction in step S007 will be described with reference to FIG. In this process, first, it is determined whether an optimization weight or an initial weight is used as a weight (S101). This determination is randomly determined so that the optimization weight is 80% and the initial weight is 20%. That is, it is determined that 8 out of 10 times use the optimization weight and 2 times use the initial weight.

  When the initial weight is used, the total display count of all advertisements distributed in rotation from the advertisement aggregation DB 23 to the current page is calculated (S102), and the total display count obtained by the calculation and the display count of each advertisement are calculated. Based on the above, the initial weight is calculated by the above formula (S103).

When using optimization weights, the site sales for the current day, the total sales for the current day, and the past sales are read and calculated from the advertisement aggregation DB 23 (S104), and the total number of display times for each advertisement is calculated (S105). Based on the above formula, the value of the advertisement is obtained, and based on this, the optimization weight of each advertisement is calculated (S106).

  Then, the distribution frequency is calculated based on one of the selected weights, and the advertisement to be distributed is selected and extracted from the advertisement registration DB 21 (S107).

  When the processing of FIG. 3 is completed and the processing returns to FIG. 2, the transmitted advertisement data is received by the content server 11 (S008), the content data provided by the content server 11 itself, the advertisement data received from the advertisement distribution server, Is generated to generate page data (S009), which is transmitted to the user terminal 40 (S010).

  The user terminal 40 receives the page data (S011) and displays it on the terminal screen (S012). In addition, although it does not show about the process of transfer, if an advertisement banner is clicked on a page, it will move to the site of the linked advertiser. When a user purchases an advertised product or service or registers as a member, the advertiser system 30 reports a result to the advertisement distribution server, and sales are generated. The advertisement distribution server records the result in the advertisement aggregation DB 23 and uses it for calculating the weight.

  In the above flowchart, steps S104 and S105 are advertisement aggregation procedures realized by the advertisement aggregation module 21a, step S106 is an optimization weight calculation procedure realized by the optimization weight calculation module 21b, and step S103 is initial weight calculation. The initial weight calculation procedure realized by the module 21c and step S107 correspond to the advertisement distribution procedure realized by the advertisement distribution module 21d.

According to the above advertisement distribution method, by determining the distribution frequency based on the optimization weight, it is possible to preferentially distribute advertisements with high sales, and the results (clicks on advertisement banners or products after clicking) The probability of occurrence of purchase) can be increased, and the advertisement reward can be increased for the advertisement distribution company. In addition, the distribution frequency of new advertisements is low only with the optimization weight, and the advertisement cannot be renewed. However, by combining the distribution frequency calculated using the initialization weight at a predetermined ratio, the increase in advertisement reward and the It is possible to achieve both the metabolism through the distribution of advertisements.

1 is a block diagram of a network system including an advertisement distribution server for realizing an advertisement distribution method according to an embodiment of the present invention. It is a flowchart which shows the whole flow of the advertisement delivery including the advertisement delivery method of embodiment. It is a flowchart which shows the mechanism of the weight calculation of the advertisement delivery method of embodiment.

Explanation of symbols

10 site 11 content server 20 advertisement distribution server 21 advertisement distribution program 21a advertisement aggregation module 21b optimization weight calculation module 21c initial weight calculation module 21d advertisement distribution module 22 advertisement registration DB
23 Advertising Total DB
30 Advertiser system 40 User terminal

Claims (4)

In a method of distributing advertisements to a public transmission source,
Aggregate sales for multiple ads delivered to a particular source to calculate the percentage of each ad's total sales,
Find the optimization weight for each ad based on the calculated percentage,
Determine the frequency of delivery of each ad by weighting it with the optimization weight,
An advertisement distribution method characterized by distributing each advertisement to the transmission source based on the determined frequency. Determining the number of times each advertisement is displayed in the transmission source, calculating the reciprocal of the proportion of the total number of advertisements distributed to the transmission source, determining the initial weight of each advertisement based on the calculated reciprocal, The advertisement distribution method according to claim 1, wherein an advertisement whose distribution frequency is determined by an optimization weight and an advertisement whose distribution frequency is determined by the initial weight are distributed at a predetermined ratio. In a program that distributes advertisements to public transmission sources,
On the computer,
An ad aggregation procedure that aggregates sales for multiple ads delivered to a specific source,
An optimization weight calculation procedure that calculates the percentage of each ad's total sales and calculates the optimization weight for each ad based on the calculated percentage,
An advertisement distribution characterized in that a frequency of distribution of each advertisement is determined by weighting with the optimization weight, and an advertisement distribution procedure for distributing each advertisement to the transmission source based on the determined frequency is executed program. The initial use for obtaining the number of times each advertisement is displayed in the transmission source, calculating the reciprocal of the proportion of the total number of advertisements distributed to the transmission source, and obtaining the initial weight of each advertisement based on the calculated reciprocal A weight calculation procedure,
4. The advertisement distribution procedure according to claim 3, wherein an advertisement whose distribution frequency is determined by the optimization weight and an advertisement whose distribution frequency is determined by the initial weight are distributed at a predetermined rate. Ad delivery program.

Images