Multi-strategy generation of product recommendations

Abstract

Techniques are described for dynamically generating recommendations for users, such as for products and other items. In at least some situations, the techniques include using multiple recommendation strategies, such as by aggregating recommendation results from multiple different recommendation strategies. Such recommendation strategies may have various forms, and may be based at least in part on data regarding prior interactions of numerous users with numerous items. In addition, information about current selections of a particular user may be gathered based at least in part on providing a GUI (“graphical user interface”) for display to the user that includes selectable information about numerous recommended items, and dynamically updating the displayed GUI with newly generated recommendations of items as the user makes selections of particular displayed recommended items (e.g., newly generated recommendations that are similar to the selected items in one or more manners, or are otherwise related to the selected items).

1 - 40 . (canceled) 
     
     
         41 . A computing system comprising:
 one or more processors; and   one or more memories having stored instructions that, when executed by at least one of the one or more processors, cause the at least one processor to obtain information about prior interactions of a plurality of consumers with a plurality of products and to provide recommendations to multiple users by, for each of the users:
 receiving, via one or more first transmissions sent over one or more computer networks to the computing system from one or more other computing systems of an online retailer, additional information about a current situation involving the user interacting with one or more first products available from the online retailer; 
 determining multiple recommendation strategies that are available to be used for the current situation; 
 selecting, based at least in part on the additional information about the current situation, only one recommendation strategy of the multiple recommendation strategies to use for the current situation; 
 automatically determining one or more products of the plurality of products to recommend to the user in the current situation by selecting the determined one or more products from product recommendations generated by the selected one recommendation strategy, the selected product recommendations being generated using information about the one or more first products and using the information about the prior interactions of the plurality of consumers with the plurality of products; and 
 providing, via one or more second transmissions sent over the one or more computer networks from the computing system to a client device of the user, information about the determined one or more products, to cause display of the provided information to the user on the client device as part of a Web page from the online retailer. 
   
     
     
         42 . The computing system of  claim 41  wherein the providing of the recommendations to the multiple users further includes, for one of the multiple users, and after the providing of the information to the client device of the one user about the determined one or more products for display as part of the Web page from the online retailer:
 receiving, via one or more third transmissions sent over the one or more computer networks from the client device of the one user to the configured computing system, an indication of a selection of at least one product of the determined one or more products by the one user via the Web page; 
 determining a group of multiple additional products of the plurality of products to recommend to the one user using a combination of the multiple recommendation strategies, the determined group of multiple additional products being based at least in part on the selected at least one product; and 
 providing, via one or more fourth transmissions sent over the one or more computer networks from the configured computing system to the client device of the one user, further information about the multiple additional products of the determined group, to cause display of the provided further information to the one user on the client device as part of an updated Web page. 
 
     
     
         43 . The computing system of  claim 42  further comprising the client device of the one user, wherein the client device is configured to receive the Web page from the online retailer, to receive the provided information about the determined one or more products from a product recommendation system on the computing system that includes the instructions stored in the one or more memories, to display the Web page to the one user with the determined one or more products being shown as part of the Web page, to receive the selection of the at least one product by the one user via the Web page, to send the indication of the selection to the product recommendation system on the computing system, to receive the provided further information about the multiple additional products of the determined group, and to, without reloading the displayed Web page, dynamically modify the displayed Web page to include information about the multiple additional products of the determined group. 
     
     
         44 . The computing system of  claim 41  wherein the providing of the recommendations to the multiple users further includes, for one of the multiple users, after the providing of the information to the client device of the one user about the determined one or more products for display to the one user, placing an order with the online retailer from the one user for one product of the determined one or more products after the one user selects that one product for purchase via one or more interactions with the Web page displayed to the one user. 
     
     
         45 - 46 . (canceled) 
     
     
         47 . The computing system of  claim 41  wherein, for each of one or more of the multiple users, the selecting of the one recommendation strategy to use for the current situation includes:
 for each of the multiple recommendation strategies, generating at least one product recommendation for the current situation based on the one or more first products and on the information about the prior interactions of the plurality of consumers with the plurality of products; and 
 identifying the selected one recommendation strategy based at least in part on comparing information about the generated at least one product recommendation for each of the multiple recommendation strategies. 
 
     
     
         48 . The computing system of  claim 41  wherein, for a first user of the multiple users, the selected one recommendation strategy to use for the current situation involving the first user is a first strategy of the multiple recommendation strategies, and for a distinct second user of the multiple users in a distinct second current situation, the selected one recommendation strategy to use for the second current situation involving the second user is a second strategy of the multiple recommendation strategies that is distinct from the first strategy. 
     
     
         49 . A computer-implemented method comprising:
 obtaining, by one or more configured computing systems of a recommendation system, information that reflects prior interactions of a plurality of users with a plurality of products;   receiving, by the one or more configured computing systems, information about a current situation involving a first user and one or more first products;   determining, by the one or more configured computing systems, one or more products of the plurality of products to recommend in the current situation to the first user based on multiple distinct recommendation strategies that are available for use in the current situation, the determining including selecting a single one recommendation strategy of the multiple recommendation strategies to use at a current time for the current situation based at least in part on the received information, and further including selecting the determined one or more products from product recommendations generated by the selected one recommendation strategy, the product recommendations being generated by the selected one recommendation strategy using information about the one or more first products and using the information about the prior interactions of the plurality of users with the plurality of products; and   providing, via one or more transmissions sent over one or more computer networks from the one or more configured computing systems to a client device of the first user, information about the determined one or more products, to cause presentation of the provided information to the first user as part of a Web page on the client device.   
     
     
         50 . The method of  claim 49  wherein the selecting of the one recommendation strategy to use at the current time for the current situation includes:
 for each of the multiple recommendation strategies, generating at least one product recommendation for the current situation based on at least one of the one or more first products of the information about the prior interactions of the plurality of users with the plurality of products; and 
 identifying the selected one recommendation strategy based at least in part on comparing information about the generated at least one product recommendation for each of the multiple recommendation strategies. 
 
     
     
         51 . The method of  claim 49  wherein the selecting of the one recommendation strategy to use at the current time for the current situation includes identifying the one recommendation strategy as being predetermined for use with the one or more first products. 
     
     
         52 . The method of  claim 49  wherein the selected one recommendation strategy generates the product recommendations from a subset of the plurality of products that are identified as popular based at least in part on the prior interactions of the plurality of users with the plurality of products. 
     
     
         53 . The method of  claim 49  wherein the selected one recommendation strategy to use at the current time for the current situation involving the first user is a first strategy of the multiple recommendation strategies, and wherein the method further comprises determining, by the one or more configured computing systems, one or more second products of the plurality of products to recommend to a second user based on selecting a second strategy of the multiple recommendation strategies that is distinct from the first strategy and based on selecting the determined one or more second products from product recommendations generated by the second strategy. 
     
     
         54 . The method of  claim 49  wherein generating of the product recommendations by the selected one recommendation strategy further uses information specific to the first user. 
     
     
         55 . The method of  claim 49  wherein generating of the product recommendations by the selected one recommendation strategy is further based at least in part on feedback from the first user related to at least one product. 
     
     
         56 . The method of  claim 49  wherein the Web page is provided to the client device of the first user from an online retailer distinct from an operator of the recommendation system, and wherein the method further comprises, after the providing of the information about the determined one or more products:
 receiving, via one or more second transmissions sent over the one or more computer networks from the client device of the first user to the one or more configured computing systems, an indication of a selection of at least one of the determined one or more products by the first user via the Web page; 
 determining a group of multiple additional products of the plurality of products to recommend to the first user using a combination of the multiple recommendation strategies, the determined group of multiple additional products being based at least in part on the selected at least one product; and 
 providing, via one or more third transmissions sent over the one or more computer networks from the one or more configured computing systems to the client device of the first user, further information about the multiple additional products of the determined group, to cause display of the provided further information to the first user on the client device as part of an updated Web page from the online retailer. 
 
     
     
         57 . The method of  claim 49  wherein the multiple recommendation strategies are each one of identifying products that are popular for a particular category of products, of identifying products that are popular for a particular retailer of products, or of identifying products with which one or more users have indicated to group the one or more first products. 
     
     
         58 . The method of  claim 49  wherein the multiple recommendation strategies include each of identifying products that are popular for a particular category of products to which the one or more first products belong, of identifying products that are popular for a particular retailer of products, of identifying products that are similar to the one or more first products based at least in part on those identified products each sharing one or more attributes with the one or more first products, of identifying products that are related to the one or more first products based at least in part on users who interact with the one or more first products in one or more manners also interacting with those identified products in the one or more manners, and of identifying products with which one or more users have indicated to group the one or more first products. 
     
     
         59 . The method of  claim 49  wherein the one or more first products correspond to one product category of multiple product categories for the plurality of products, and wherein the selected one recommendation strategy includes at least one of selecting products that are frequently purchased from the one product category or of selecting products that are frequently viewed from the one product category. 
     
     
         60 . The method of  claim 49  wherein the one or more first products correspond to at least one search result of multiple search results of a search initiated by the first user, and wherein the selected one recommendation strategy includes selecting products that are related to the one or more first products in the prior interactions of the plurality of users. 
     
     
         61 . The method of  claim 49  wherein the one or more first products are indicated by the first user, and wherein the selected one recommendation strategy includes selecting products that are related to the one or more first products in the prior interactions of the plurality of users. 
     
     
         62 . The method of  claim 49  wherein the plurality of users are customers of an online retailer, wherein the information that reflects the prior interactions of the plurality of users with the plurality of products is obtained based at least in part on tracking interactions of the customers with the online retailer, wherein the recommendation systems provides a multi-strategy product recommendation service that determines product recommendations for multiple customer users of the online retailer, and wherein the method further comprises automatically placing an order with the online retailer for the first user of one product of the determined one or more products after the first user selects that one product for purchase via one or more interactions with the provided information about the determined one or more products. 
     
     
         63 . A non-transitory computer-readable medium having stored contents that configure a computing device to perform a method, the method comprising:
 obtaining, by the configured computing device, information that reflects prior interactions of a plurality of users with a plurality of items;   receiving, by the configured computing device, information about a first user and one or more first items;   determining, by the configured computing device, one or more items of the plurality of items to recommend to the first user based on multiple distinct recommendation strategies, the determining including selecting one recommendation strategy of the multiple recommendation strategies to use to currently generate item recommendations for the first user based at least in part on the received information, and the determining further including selecting the determined one or more items from the currently generated item recommendations by the selected one recommendation strategy, the item recommendations being generated by the selected one recommendation strategy using the information about the prior interactions of the plurality of users with the plurality of items; and   providing, via one or more transmissions sent over one or more computer networks from the configured computing device to a client device of the first user, information about the determined one or more items, to cause presentation of the provided information to the first user on the client device.   
     
     
         64 . The non-transitory computer-readable medium of  claim 63  wherein the one or more first items are products, and wherein generating of the item recommendations by the selected one recommendation strategy further uses information about the one or more first items. 
     
     
         65 . The non-transitory computer-readable medium of  claim 63  wherein generating of the item recommendations by the selected one recommendation strategy further uses information about the first user. 
     
     
         66 . The non-transitory computer-readable medium of  claim 63  wherein the selecting of the one recommendation strategy to use includes:
 for each of the multiple recommendation strategies, generating at least one item recommendation based on at least one of the one or more first items or of the information about the prior interactions of the plurality of users with the plurality of items; and 
 identifying the selected one recommendation strategy for use in a current situation involving the first user and the one or more first items based at least in part on comparing information about the generated at least one item recommendation for each of the multiple recommendation strategies and by determining to use only the selected one recommendation strategy for the current situation, so that all other recommendation strategies of the multiple recommendation strategies are excluded from use in the current situation. 
 
     
     
         67 . The non-transitory computer-readable medium of  claim 63  wherein the multiple recommendation strategies are each one of identifying items that are popular for a particular category of items, of identifying items that are popular for a particular retailer, or of identifying items with which one or more users have indicated to group the one or more first items. 
     
     
         68 . The non-transitory computer-readable medium of  claim 63  wherein the multiple recommendation strategies include each of identifying items that are popular for a particular category of items to which the one or more first items belong, of identifying items that are popular for a particular retailer, of identifying items that are similar to the one or more first items based at least in part on those identified items each sharing one or more attributes with the one or more first items, of identifying items that are related to the one or more first items based at least in part on users who interact with the one or more first items in one or more manners also interacting with those identified items in the one or more manners, and of identifying items with which one or more users have indicated to group the one or more first items. 
     
     
         69 . The non-transitory computer-readable medium of  claim 63  wherein the plurality of users are customers of an online retailer, wherein the information that reflects the prior interactions of the plurality of users with the plurality of products is obtained based at least in part on tracking interactions of the plurality of customers with the online retailer, wherein the presentation of the provided information to the first user on the client device involves adding the provided information to a Web page supplied from the online retailer to the client device, and wherein the configured computing device is part of a multi-strategy product recommendation service that determines item recommendations for multiple customer users of the online retailer. 
     
     
         70 . The non-transitory computer-readable medium of  claim 63  wherein the computer-readable medium is a memory of the configured computing device, and wherein the stored contents are software instructions that, when executed, program the configured computing device to perform the method.