Personalized web content distribution system

Abstract

A personalized web content distribution system comprises storage (12) for home webpages, in which each webpage is associated firstly with a segment identifier and secondly with a webpage address, some of the webpages being associated with the same webpage address such that each of these webpages defines a personalized version of a generic webpage associated with said same webpage address, a web server (4) designed to receive a generic request to access a webpage, said request being sent from a device (3) connected to the Internet, comprising a webpage address provided with a fingerprint of said device, to determine a segment identifier associated with said fingerprint, and to access said storage (12) in order to obtain and return, to said device (3), a webpage associated with said webpage address of the generic request and with said segment identifier.

Description

[0001] Description

[0002] Title: Personalized Internet Content Distribution System

[0003] The invention relates to web servers and more particularly to the provision of personalized content.

[0004] It has now become common practice to personalize website content based on visitors. This personalization makes content more engaging and generally improves the user experience.

[0005] Currently, there are two families of personalization methods:

[0006] - methods based on prior identification, and

[0007] - agnostic methods.

[0008] In the case of identification-based methods, a visitor has an account on the web server they are visiting, and the web server is therefore able to offer personalized content for the user, since it knows who the user is.

[0009] These pre-identification methods lack flexibility and fluidity for the user. Indeed, a web page requires a user to first authenticate themselves to access personalized content, which harms their experience by adding a step that also requires them to remember or store a username and password.

[0010] The second family of methods, namely that of agnostic methods, classically uses two types of solutions.

[0011] The first type of solution consists of displaying a classic web page, then personalizing it as soon as possible. This personalization is generally achieved either by automatic reloading, unsolicited by the user, as soon as this user has been associated with a user cluster, or by personalizing subsets of the page using JavaScript code. Reloading the page is not compatible with a satisfactory user experience (UX) and can even be perceived as a defect of the site. The use of JavaScript poses problems in keeping the code up to date, as well as the impossibility of implementation in secure environments, since JavaScript execution is often blocked.

[0012] Additionally, the fact that the user sees the customization happening right in front of them can make them aware of the customization, and can give them the impression that it is artificial, or even intended to manipulate them. This significantly degrades the UX.

[0013] The invention improves the situation. To this end, it proposes a system for distributing personalized Internet content comprising a storage (12) of home web pages, in which each web page is associated on the one hand with a segment identifier and on the other hand with a web page address, some of the web pages being associated with the same web page address such that each of these web pages defines a personalized version of a generic web page associated with said same web page address, a web server arranged to receive a generic request for access to a web page, said request being issued from a device connected to the Internet, comprising a web page address provided with a fingerprint of said device, to determine a segment identifier associated with said fingerprint, and to access said storage to obtain and return to said device a web page associated with said web page address of the generic request and said segment identifier.

[0014] This system is particularly advantageous because it allows for customization as soon as the web page loads. Furthermore, since the user is accessing a web page that is already personalized, which is therefore more relevant, it improves its UX without them realizing it.

[0015] According to various embodiments, the invention may have one or more of the following characteristics:

[0016] - the web server includes a reverse proxy arranged to query a database in which each fingerprint is associated with a segment identifier,

[0017] - the reverse proxy is an NGINX server, - the reverse proxy is a microservice,

[0018] - the web server is arranged to receive a request for access to a web page sent by a user from a device connected to the Internet, without a fingerprint of said device, and to generate said fingerprint and return it with a general version of said web page, and

[0019] - the web server is arranged to receive a generic request for access to a web page comprising a web page address provided with a segment identifier and to directly access said storage with this generic access request.

[0020] The invention also relates to a method for distributing personalized Internet content comprising the following operations: a) receiving a generic request for access to a web page, said request being issued by a user from a device connected to the Internet, comprising a web page address provided with a fingerprint of said device, b) determining a segment identifier associated with said user from said fingerprint of said device, c) accessing a storage of home web pages, in which each web page is associated on the one hand with a segment identifier and on the other hand with a web page address, some of the home web pages being associated with the same web page address such that each of these home web pages defines a personalized version of a generic web page associated with said same web page address,and retrieving a web page associated with said web page address of the generic request and said segment identifier determined in operation b), and d) returning to said device the web page retrieved in operation c).,

[0021] Other characteristics and advantages of the invention will appear more clearly on reading the following description, taken from examples given for illustrative and non-limiting purposes, taken from the drawings in which:

[0022] [Fig.l] represents a generic diagram of a first embodiment of a personalized Internet content distribution system. The drawings and the description below contain, for the most part, elements of a certain character. They may therefore not only serve to better understand the present invention, but also contribute to its definition, if necessary.

[0023] Figure 1 represents a generic diagram of a personalized Internet content distribution system 2 according to the invention. The personalized Internet content distribution system 2 receives requests for access to web contents, in particular web pages from one or more devices 3 which are connected to the Internet.

[0024] In the example described here, the personalized Internet content distribution system 2 comprises one or more web servers 4 including a reverse proxy 6 and one or more servers 8, a database 10 associating Reqid request identifiers and SN segment identifiers, and a storage 12 which contains home web pages.

[0025] In the example described here, the reverse proxy 6 and the server(s) 8 may be implemented in the form of appropriate computer code or program executed on one or more processors. Processors should be understood to mean any processor suitable for the calculations described below. Such a processor may be implemented in any known manner, in the form of a microprocessor for a personal computer, a dedicated chip of the FPGA or SoC type, a computing resource on a grid or in the cloud, a cluster of graphics processing units (GPUs), a microcontroller, or any other form suitable for providing the computing power necessary for the implementation described below. One or more of these elements may also be implemented in the form of specialized electronic circuits such as an ASIC. A combination of processor and electronic circuits may also be considered. Processors dedicated to machine learning may also be considered.The reverse proxy 6 and the server(s) 8 will be described in more detail below and could be combined as a single unit. In addition, they include means of communication via the Internet allowing them to receive requests for Internet content from one or more of the devices 8, and to return this content to them if it is accessible to them. In the example described here, the database 10 and the storage 12 can be implemented on any type of data storage suitable for receiving digital data: hard disk, flash memory hard disk, flash memory in any form, RAM, magnetic disk, locally or cloud distributed storage, etc. The data calculated by the system 2 can be stored on any type of memory similar to the memory of the database 10 and the storage 12, or on them. This data can be erased after the system 2 has performed its tasks or retained.

[0026] The web server 4 has the conventional function of receiving requests for Internet content from the device 8 and returning this content to the latter. In a conventional manner according to the http (or https) protocol, this is achieved by issuing a GET request which contains the web address or uniform resource locator (URL) which a user of the device 8 seeks to access.

[0027] In classic 4 web servers, it is not possible to personalize a web resource without the user being authenticated, that is, without having opened an identification session in some form with the web server. It is this authentication that allows the web server to choose personalized content for the user from the same generic web page address. Personalized content means, for example, displaying a page on which all or part of the content is modified according to, for example, the desires and / or tastes and / or preferences and / or habits and / or location and / or the language of the authenticated user. For example, on a sports news site, this can be achieved by highlighting recent content relating to one or more teams that a user has indicated they prefer.

[0028] The personalized Internet content distribution system 2 according to the invention is particular in that it allows specific processing of home page requests in the absence of user authentication. In other words, the personalized Internet content distribution system 2 receives a generic request without a user identifier, and is nevertheless arranged to provide this user with a version of the web page which is adapted to his desires and / or tastes and / or preferences. This is all the more remarkable since the web page is a special case where the web server has no access to any information which allows the user experience to be personalized in the absence of authentication, whereas the web server can try to implement personalization for the other pages of a site, for example on the basis of the path which allowed access to them from the web page.

[0029] To this end, the Applicant has discovered that it is advantageous to pre-process the request to access the web page, even if this means delaying its loading on the client side by a few milliseconds, in order to provide a pre-personalized web page.

[0030] Thus, the web server 4 receives a request to access a web page Hp.html referenced GET (Hp. html) in figure 1, and transmits it directly to a reverse proxy 6. The use of reverse proxies 6 is known in itself mainly in order to achieve load balancing for web servers 4 comprising a large number of servers 8 making it possible to manage large flows of visitors.

[0031] In the embodiment described here, the reverse proxy 6 is partially diverted from its primary use in order to allow the web page returned to the device 3 to be personalized. For this, the reverse proxy 6 uses a Reqid identifier of the device 3. This identifier, when it exists, is stored locally in a cookie on the device 3, and is transmitted to the web server 4 with the GET(Hp.html) request.

[0032] In the case where the device 3 does not contain a Reqid identifier, it is generated and stored for the next visit. In the example described here, the Reqid identifier is generated by a fingerprint.js package (https: / / fingerprintjs.com), which guarantees to obtain a unique identifier for each device 3. Alternatively, other methods could be used to generate the Reqid identifier.

[0033] From the Reqid identifier, the reverse proxy 6 queries the database 10 to retrieve a segment identifier. The segment identifiers are used to choose personalized content. More specifically, the users of the devices 3 are grouped according to their behavior. For example, they may be grouped because they have visited a particular page previously, or a succession of particular pages, or because their behavior has been similar (e.g., putting an item in a shopping cart but not purchasing, selecting particular information, etc.), etc.

[0034] Once this grouping is determined, a personalized web page is defined for this group of users, also called a segment. Finally, the database 10 is updated to associate on the one hand the Reqid identifiers and on the other hand the segment identifiers that correspond to them.

[0035] Thus, when the reverse proxy 6 accesses the database 10, it retrieves a segment identifier Sn which designates a group of users for which a particular personalization has been planned.

[0036] In order to optimize performance, the database 10 may advantageously be implemented in the form of a cache within the reverse proxy 6. Alternatively, it may be implemented in any other useful manner.

[0037] Then, the reverse proxy 6 uses the web page - segment identifier pair to obtain the customized version of the web page Hp.html for the segment identifier Sn in the storage 12. In the example described here, this is achieved by appending a suffix corresponding to the segment identifier Sn to the name of the web page. Other implementations will be possible, such as storage in directories associated with the segment identifier, other naming conventions, etc.

[0038] In the example described here, the reverse proxy 6 is an NGINX server. Alternatively, the reverse proxy 6 could be another reverse proxy, such as a HAProxy or Traefik server. Still alternatively, the reverse proxy 6 could be a microservice, for example implemented using a lambda server from the company Amazon Web Services. Alternatively, the device 3 can store the segment identifier Sn in the same way that it stores the identifier Reqid once the latter has been determined. Thus, when the web server 4 receives a request to access the page Hp.html, the segment identifier Sn is transmitted like other cookies, and the web server 4 can directly request the storage 12 with the URL of the web page Hp.html and the segment identifier Sn in order to return the personalized version.

[0039] Thus, the server 8 can return to the device 3 the page Hp_Sn.html which is the personalized version of the page Hp.html for the segment identifier segment Sn associated with the device 3. Alternatively, the storage 12 could contain not an entire copy of the personalized web page, but only personalization elements, the server 8 then being responsible for combining them with the generic version of the web page before sending it to the device 3.

[0040] The personalized Internet content distribution system 2 therefore makes it possible to provide the user of the device 3 with a personalized version of the web page of a site, without the user being authenticated, and without the user being able to realize that a personalization has been implemented between his request and the display on his screen of the web page. In the tests implemented by the Applicant, the latter was able to verify that the implementation of the system 2 only requires 10 ms, which makes this execution time transparent for the user.

Claims

Demands

1. A personalized Internet content distribution system comprising a storage (12) of home web pages, in which each web page is associated on the one hand with a segment identifier and on the other hand with a web page address, some of the web pages being associated with the same web page address such that each of these web pages defines a personalized version of a generic web page associated with said same web page address, a web server (4) arranged to receive a generic request for access to a web page, said request being issued from a device (3) connected to the Internet, comprising a web page address with a fingerprint of said device, to determine a segment identifier associated with said fingerprint, and to access said storage (12) to obtain and return to said device (3) a web page associated with said web page address of the generic request and said segment identifier.

2. A personalized Internet content distribution system according to claim 1, wherein the web server (4) includes a reverse proxy (6) arranged to query a database (10) in which each fingerprint is associated with a segment identifier.

3. A personalized Internet content distribution system according to claim 2, wherein the reverse proxy (6) is an NGINX server.

4. A personalized Internet content distribution system according to claim 2, wherein the reverse proxy (6) is a microservice.

5. A system according to any one of the preceding claims, wherein the web server (4) is arranged to receive a request to access a web page issued by a user from a device (3) connected to the Internet, devoid of a fingerprint of said device (3), and to generate said fingerprint and return it with a generic version of said web page.

6. A system according to any one of the preceding claims, wherein the web server (4) is arranged to receive a generic web page access request comprising a web page address equipped with a segment identifier and to access said storage (12) directly with this generic access request.

7. A method for distributing personalized Internet content comprising the following operations: a) receiving a generic request to access a web page, said request being issued by a user from an Internet-connected device, comprising a web page address with a fingerprint of said device, b) determining a segment identifier associated with said user from said fingerprint of said device, c) accessing a storage of home web pages, in which each web page is associated on the one hand with a segment identifier and on the other hand with a web page address, some of the home web pages being associated with the same web page address such that each of these home web pages defines a personalized version of a generic web page associated with said same web page address, and retrieving a web page associated with said web page address from the generic request and said segment identifier determined in operation b),and d) return to said device the web page retrieved from operation c).

Images