Systems and Methods for Recommended Content Platform

• US 20130036191A1
• Filed: 10/31/2011
• Published: 02/07/2013
• Est. Priority Date: 06/30/2010
• Status: Active Grant

• Current Assignee: Leaf Group Ltd. (Graham Holdings Company)
• Original Assignee: Demand Media Incorporated (Graham Holdings Company)
• Inventors: Fink, Kevin, Otis, Todd
• Granted Patent: US 9,721,035 B2
• Time in Patent Office: Days
• US Class Current: 709/217
• CPC Class Codes: G06F 16/9577 Optimising the visualization of content, e.g. distillation of HTML documents

Abstract

Systems and methods for presentation of content, or a title or link to content or presentation to a user on webpages of a website are provided. In one embodiment, a recommended content platform optimizes the links displayed on web pages based on a portfolio of content sources which determine which links are displayed in a given page view. The algorithms may or may not utilize data relating to user interaction with webpages and may produce different sets of links depending on the content of a webpage and a position on a webpage where the links are to be displayed. A given webpage may present links from multiple content sources displayed using multiple user interface elements distributed throughout the webpage.

Claims

1. A method comprising:

receiving, over a network, a request for recommended content from a requester, the request comprising a content unit, the content unit being associated with a respective plurality of content sources, each respective content source comprising a respective content recommendation methodology and at least one respective data source, wherein the respective content source identifies links to recommended content items by applying the respective content recommendation methodology to the at least one respective data source;

selecting, using a computing device, one content source of the plurality of content sources;

identifying, using the computing device, a plurality of links to recommended content items accessible over the network using the one content source of the plurality of content sources; and

transmitting, over the network, the plurality of links to recommended content items to the requester.

2. The method of claim 1 wherein the one content source of the plurality of content sources is selected randomly.

3. The method of claim 2 wherein each content source of the plurality of content sources further comprises a respective weight, wherein the selection of the one content algorithm of the plurality of content algorithms is biased by the respective weights of each content source of the plurality of content algorithms.

4. The method of claim 1 wherein each content source of the plurality of content sources further comprises a respective weight, wherein the content source having the greatest weight is selected.

5. The method of claim 1 wherein each content source of the plurality of content sources further comprises a respective weight that varies based on a value of an input parameter, wherein the request for recommended content additionally comprises the input parameter, wherein the content source having the greatest weight given the value of the input parameter is selected.

6. The method of claim 5 wherein the input parameter is selected from the list:

installation, category, category and subcategory, and a keyword.

7. The method of claim 1 wherein the one content source of the plurality of content sources further comprises a minimum number of links, such that where the one algorithm identifies a number of links to recommended content items less than the minimum number of links, a default content source is used to identify additional links to recommended content items that are added to the plurality of links to recommended content items, such that a number of links recommended content items in the plurality of links to recommended content items is equal to the minimum number of links.

8. The method of claim 1 wherein the one content source of the plurality of content sources additionally comprises a maximum number of links, such that where the one content source identifies a number of links to recommended content items greater than the maximum number of links, a subset of links to recommended content is randomly selected from the plurality of links to recommended content, such that a number of links to recommended content in the subset of links to recommended content is transmitted to the requester.

9. The method of claim 1 wherein the one content source of the plurality of content sources additionally comprises a maximum number of links, such that where the one content source identifies a number of links to recommended content items greater than the maximum number of links, a subset of links to recommended content is selected from the plurality of links to recommended content by truncating the plurality of links, such that a number of links to recommended content in the subset of links to recommended content is transmitted to the requester.

10. The method of claim 1 wherein the one content source of the plurality of content sources is a dependant algorithm, such that the respective at least one dataset associated with the one algorithm comprises data relating to user interaction with at least one webpage and the respective content recommendation methodology associated with the one algorithm identifies links to recommended content items based, at least in part, the data relating to user interaction with at least one webpage.

11. The method of claim 10 wherein the data relating to user interaction with at least one webpage comprises at least one of:

click-through-rate (CTR), subsequent-views-per-session (PV), ad clicks, ad clicks per session, revenue, revenue per session, and weighted views.

12. The method of claim 1 wherein the one algorithm of the plurality of content sources is an independent algorithm, such that the respective content recommendation methodology associated with the one algorithm identifies links to recommended content items based on data that does not include data relating to user interaction with webpages.

13. The method of claim 1 wherein the request additionally comprises an identification of a webpage, and wherein one algorithm of the plurality of content sources is an attached algorithm, such that the respective content recommendation methodology associated with the one algorithm identifies links to recommended content items based on, at least in part, the identification of a webpage.

14. The method of claim 1 wherein the request additionally comprises an identification of a webpage, and wherein one algorithm of the plurality of content sources is an detached algorithm, such that the respective content recommendation methodology associated with the one algorithm does not identify links to recommended content items using the identification of a webpage.

15. The method of claim 1 wherein the request additionally comprises an order for links to recommended content, wherein the plurality of links to recommended content items is transmitted, over the network, to the requester in the order for links to recommended content.

16. A method comprising:

receiving, over a network, a request for recommended content from a requester, the request comprising a content source, the content source comprising a plurality of content sources, each respective content source comprising a respective content recommendation methodology and at least one respective data source, wherein the respective content source identifies links to recommended content items by applying the respective content recommendation methodology to the at least one respective data source;

for each content source of the plurality of content sources,

identifying, using the computing device, a respective plurality of links to recommended content items accessible over the network using the respective content source,

aggregating, using the computing device, a respective plurality of links into a aggregated plurality of links to recommended content items; and

transmitting, over the network, the aggregated plurality of links to recommended content items.

17. The method of claim 16 wherein each content source of the plurality of content sources further comprises a respective maximum number of links, such that where a respective content source identifies a number of links to recommended content items greater than the respective maximum number of links, a respective subset of links to recommended content is randomly selected from the respective plurality of links to recommended content, such that a number of links to recommended content in the subset of links to recommended content is aggregated into the aggregated plurality of links, and wherein each content source of the plurality of content sources further comprises a minimum number of links, such that where a respective content source identifies a number of links to recommended content items less than the respective minimum number of links, a default content source is used to identify additional links to recommended content items that are added to the respective plurality of links to recommended content items, such that a number of links recommended content items in the respective plurality of links to recommended content items is equal to the respective minimum number of links.

18. The method of claim 16 wherein:

a first content source of the plurality of content sources is a dependant algorithm, such that the respective at least one dataset associated with the first content source comprises data relating to user interaction with at least one webpage and the respective content recommendation methodology associated with the first content source identifies links to recommended content items based, at least in part, the data relating to user interaction with at least one webpage, and wherein

a second content source of the plurality of content sources is an independent algorithm, such that the respective content recommendation methodology associated with the second content source identifies links to recommended content items based on data that does not include data relating to user interaction with webpages.

19. A machine readable media embodying instructions for a widget embedded in a webpage, the instructions causing a data processing system to performs a method, the method comprising:

identifying a plurality of content units within the webpage, each content unit of the plurality of content units comprising a respective definition of a user interface element and a respective content source for retrieving links to recommended content from a server;

generating a request for recommended content, the request comprising, for each content unit of the plurality of content units, an identification of the respective content unit and the respective content source associated with the respective content unit;

transmitting, over a network, the request for recommended content to the server;

receiving, over the network, a plurality of links to recommended content items comprising at least one respective link for each content unit of the plurality of content units;

generating, for each of the plurality of content units, a respective one user interface element of a plurality of user interface elements using the respective definition of a user interface element of the respective content unit of the plurality of content units, wherein the respective one user interface element is populated with the respective at least one respective link for the respective content unit;

displaying each user interface element of the plurality of user interface elements on a display device in association with a display of the webpage.

20. A computer system comprising:

a storage device storing source code for a plurality of webpages;

a memory; and

at least one processor coupled to the memory to;

for each webpage of the plurality of webpages,

identify a respective plurality of content units within the respective source code for the respective webpage, each content unit of the respective plurality of content units comprising a respective definition of a user interface element and a respective content source for retrieving links to recommended content from a server;

generate a request for recommended content, the request comprising, for each content unit of the respective plurality of content units, an identification of the respective content unit and the respective content source associated with the respective content unit;

transmit, over a network, the request for recommended content to the server;

receive, over the network, a plurality of links to recommended content items comprising at least one respective link for each content unit of the respective plurality of content units;

generate, for each of the respective plurality of content units, a respective one user interface element of a respective plurality of user interface elements using the respective definition of a user interface element of the respective content unit of the plurality of content units, wherein the respective one user interface element is populated with the respective at least one at least one respective link for the respective content unit;

update the respective source code for the respective webpage on the storage device to insert the respective plurality of user interface elements into the respective source code.