Determining search query specificity
View Patent ↗Systems and methods described and suggested herein utilize various techniques to assess the degree of specificity of search queries. By assessing the degree of specificity of search queries, the systems and methods may be used to determine where the submitting users' intentions fall along the continuum between specific and general searching. Thus, the systems and methods may be used to, among other uses, deliver search results that more closely coincide with the users' intentions. For example, if data associated with a search query indicates that users typically submit the search query with the intention of specific searching, embodiments may tailor search results to include a limited number of highly relevant search-result items. On the other hand, for example, if data associated with a search query indicates that users typically submit the search query with the intention of general searching, embodiments may tailor search results to include a diverse set of search-result items from multiple categories.
1. A computer-implemented method of assessing specificity of a search query, the method comprising:
under the control of one or more computer systems configured with executable instructions,
obtaining clickstream data associated with the search query, the clickstream data comprising:
indications of items listed in search results presented to users in response to the users submitting the search query in a search engine;
for each of the items, an indication of a number of times the item was listed in the search results and an indication of a number of first-clicks the item received, wherein a first-click occurs when a user clicks, previews, hovers, or otherwise interacts with an item first in time before the other items listed in the search results;
for each of the items, determining a first-click probability value by dividing the number of times the users first-clicked the item by a total number of first-clicks received across all of the items listed in the search results; and
determining a specificity score for the search query based at least in part on the first-click probability values of the items listed in the search results, wherein the specificity score for the search query indicates a uniformity of a distribution of first-clicks across the items presented in the search results, and wherein the distribution is indicative of how the first-clicks are spread across the items presented in the search results, wherein the more evenly distributed the distribution of first-clicks, the lower the specificity score and the more likely the search query is submitted by users with the intention of general searching, and wherein the more skewed the distribution of first-clicks, the higher the specificity score and the more likely the search query is submitted by users with the intention of specific searching.
2. The computer-implemented method of claim 1 , further comprising:
determining a first-click entropy for the search query based at least in part on the first-click probability values of the items listed in the search results presented to the users in response to the users submitting the search query, wherein the specificity score is based at least in part on the first-click entropy.
3. The computer-implemented method of claim 1 , wherein the more evenly the first-click probability values are distributed across the items listed in the search results, the lower the specificity score and the more likely the search query is submitted by users with the intention of general searching.
4. The computer-implemented method of claim 1 , wherein the more skewed the first-click probability values are toward one of the items listed in the search results, the higher the specificity score and the more likely the search query is submitted by users with the intention of specific searching.
5. The computer-implemented method of claim 1 , further comprising:
comparing the specificity score to a threshold specificity score;
determining that users likely submit the search query with the intention of specific searching if the specificity score is equal to or greater than the threshold specificity score; and
determining that users likely submit the search query with the intention of general searching if the specificity score is less than the threshold specificity score.
6. A computer-implemented method of assessing specificity of a search query, the method comprising:
under the control of one or more computer systems configured with executable instructions,
obtaining data associated with the search query, the data comprising indications of search-result items and, for each search-result item, a corresponding number of first-clicks, the search-result items being items presented to users in response to the users submitting the search query, the number of first-clicks indicating the number of times the users selected the corresponding search-result item first in time before other items;
using the indications of the search-result items and the corresponding number of first-clicks to create a distribution of first-clicks across the search result items, wherein the distribution is indicative of how the first-clicks are spread across the search-result items; and
determining a specificity score for the search query based at least in part on the distribution of first-clicks across the search result items, wherein the more evenly distributed the distribution of first-clicks, the lower the specificity score and the more likely the search query is submitted by users with the intention of general searching, and wherein the more skewed the distribution of first-clicks, the higher the specificity score and the more likely the search query is submitted by users with the intention of specific searching.
7. The computer-implemented method of claim 6 , wherein using the indications of the search-result items and the corresponding number of first-clicks to create the distribution of first-clicks across the search result items, comprises:
determining a first-click probability value for each of the search-result items by dividing the number of first-clicks for the search-result item by a total number of first-clicks received across all of the search-result items.
8. The computer-implemented method of claim 7 , further comprising:
comparing the specificity score to a threshold specificity score;
determining that users likely submit the search query with the intention of specific searching if the specificity score is equal to or greater than the threshold specificity score; and
determining that users likely submit the search query with the intention of general searching if the specificity score is less than the threshold specificity score.
9. The computer-implemented method of claim 8 , further comprising:
if the specificity score is equal to or above the threshold and users likely submit the search query with the intention of specific searching, highlighting the search-result item that has the highest first-click probability value so that the users can easily identify the item when reviewing the search-result items.
10. The computer-implemented method of claim 8 , further comprising:
if the specificity score is equal to or above the threshold and users likely submit the search query with the intention of specific searching, presenting one or more sponsored links related to the search-result item that has the highest first-click probability value.
11. The computer-implemented method of claim 8 , further comprising:
if the specificity score is equal to or above the threshold and users likely submit the search query with the intention of specific searching, directing users to a webpage associated with the search-result item that has the highest first-click probability value.
12. The computer-implemented method of claim 8 , further comprising:
if the specificity score is below the threshold and users likely submit the search query with the intention of general searching, presenting the users with search-result items from a plurality of categories that may be relevant to the search query.
13. The computer-implemented method of claim 12 , wherein a category label is associated with each the search-result items.
14. The computer-implemented method of claim 8 , further comprising:
if the specificity score is below the threshold and users likely submit the search query with the intention of general searching, presenting the users with tools to help narrow down the search-result items.
15. A computer-implemented method, comprising:
under the control of one or more computer systems configured with executable instructions,
obtaining data associated with a search query, the data comprising:
an indication of the number of times the users submitted the search query; and
indications of search-result items and corresponding number of first-clicks, the search-result items being items presented to the users in response to the users submitting the search query, the number of first-clicks being the number of times the users selected the corresponding search-result item first in time before other items; and
determining a first-click probability value for each of the search-result items based at least in part on a relationship between the number of first-clicks for the search-result item and a total number of first-clicks received across all of the search-result items;
using the first-click probability values of the search-result items to create a distribution of first-clicks across the search result items, wherein the distribution is indicative of how the first-clicks are spread across the search-result items;
determining a specificity score for the search query based at least in part on the distribution of first-clicks across the search result items;
responsive to the users submitting the search query, tailoring search results presented to the users based at least in part on the specificity score, wherein the more evenly distributed the distribution of first-clicks, the lower the specificity score and the more likely the search query is submitted by users with the intention of general searching, and wherein the more skewed the distribution of first-clicks, the higher the specificity score and the more likely the search query is submitted by users with the intention of specific searching.
16. The computer-implemented method of claim 15 , wherein tailoring search results comprises:
if the specificity score is equal to or above a threshold, highlighting the search-result item that has the highest first-click probability value so that the users can easily identify the item when reviewing the search-result items.
17. The computer-implemented method of claim 15 , wherein tailoring search results comprises:
if the specificity score is above a threshold, presenting one or more sponsored links related to the search-result item that has the highest first-click probability value.
18. The computer-implemented method of claim 15 , wherein tailoring search results comprises:
if the specificity score is above a threshold, directing users to a webpage associated with the search-result item that has the highest first-click probability value.
19. The computer-implemented method of claim 15 , wherein tailoring search results comprises:
if the specificity score is below a threshold, presenting the users with search-result items from a plurality of categories that may be relevant to the search query.
20. A computer program product embedded in a computer-readable storage medium, comprising:
program code for obtaining data associated with a search query, the data
comprising indications of search-result items and corresponding number of first-clicks, the search-result items being items presented to users in response to the users submitting the search query, the number of first-clicks being the number of times the users selected the corresponding search-result item first in time before other items;
program code for using the indications of the search-result items and the corresponding number of first-clicks to create a distribution of first-clicks across the search result items, wherein the distribution is indicative of how the first-clicks are spread across the search result items; and
program code for determining a specificity score for the search query based at least in part on the distribution of first-clicks across the search result items, wherein the more evenly distributed the distribution of first-clicks, the lower the specificity score and the more likely the search query is submitted by users with the intention of general searching, and wherein the more skewed the distribution of first-clicks, the higher the specificity score and the more likely the search query is submitted by users with the intention of specific searching.
21. The computer program product of claim 20 , wherein using the indications of the search-result items and the corresponding number of first-clicks to create the distribution of first-clicks across the search result items, comprises:
determining a first-click probability value for each of the search-result items by dividing the number of first-clicks for the search-result item by a total number of first-clicks received across all of the search-result items.
22. A system, comprising:
one or more processors; and
one or more memory device that collectively include instructions that, when executed by the processor, cause the processor to:
obtain data associated with a search query, the data comprising indications of search-result items and corresponding number of first-clicks, the search-result items being items presented to users in response to the users submitting the search query, the number of first-clicks being the number of times the users selected the corresponding search-result item first in time before other items;
use the indications of the search-result items and the corresponding number of first-clicks to create a distribution of first-clicks across the search result items, wherein the distribution is indicative of how the first-clicks are spread across the search-result items;
determine a specificity score for the search query based at least in part on the distribution of first-clicks across the search result items, wherein the more evenly distributed the distribution of first-clicks, the lower the specificity score and the more likely the search query is submitted by users with the intention of general searching, and wherein the more skewed the distribution of first-clicks, the higher the specificity score and the more likely the search query is submitted by users with the intention of specific searching;
use the specificity score to select search results; and
provide the content to the user.
23. The system of claim 22 , the instructions that, when executed by the processor, further cause the processor to:
determine a first-click probability value for each of the search-result items by dividing the number of first-clicks for the search-result item by a total number of first-clicks received across all of the search-result items.