Determining search relevance from user feedback

A method, comprising:

• obtaining, via at least one of one or more computing devices, a plurality of relevance indications corresponding to an indication of a respective relevance of individual ones of a first plurality of search results corresponding to a search query, wherein the respective relevance is associated with a relevance score based at least in part upon search behavior data associated with previous submissions of the search query;
• determining, via at least one of the one or more computing devices, whether individual ones of the plurality of relevance indications meet a relevance threshold, where the relevance threshold is set dynamically as a point at which relevance scores associated with the first plurality of search results begin to decline by a predetermined number of score units;
• performing, via at least one of the one or more computing devices, a subsequent search to produce a second plurality of search results;
• excluding, via at least one of the one or more computing devices, search results that have a relevance score that fails to meet the relevance threshold from the second plurality of search results;
• receiving, via at least one of the one or more computing devices, a sort criterion from a client device, wherein the sort criterion comprises at least one of a price of items listed in a product catalog or a popularity of the items listed in the product catalog;
• sorting, via at least one of the one or more computing devices, the second plurality of search results according to the sort criterion;
• and
• providing, via at least one of the one or more computing devices, the sorted second plurality of search results to the client device.

A system, comprising:

• at least one computing device comprising a processor and a memory;
• and
• an application executable in the at least one computing device, the application comprising:
• logic that, in response to receiving a search query in a first search request from a client device, performs a search for the search query to produce a first plurality of search results, the search being associated with a product catalog and the first plurality of search results corresponding to items listed in the product catalog;
• logic that provides the first plurality of search results corresponding to the search query to the client device;
• logic that receives a plurality of relevance indications corresponding to an indication of a respective relevance of at least one of the first plurality of search results, wherein the respective relevance is associated with a relevance score based at least in part upon search behavior data associated with previous submissions of the search query;
• logic that determines whether individual ones of the plurality of relevance indications meet a relevance threshold, the relevance threshold set dynamically as a point at which relevance scores associated with the first plurality of search results begin to decline by a predetermined number of score units;
• logic that performs a subsequent search for the search query in response to a second search request to produce a second plurality of search results;
• logic that excludes search results that have a relevance score that fails to meet the relevance threshold from the second plurality of search results;
• logic that receives a sort criterion from the client device, wherein the sort criterion comprises at least one of a price of items listed in the product catalog or a popularity of the items listed in the product catalog;
• logic that sorts the second plurality of search results according to the sort criterion;
• and
• logic that provides the sorted second plurality of search results to the client device.