Skip Navigation
Expand
Knowledge Advanced Product and Category Filtering/Faceting Behavior for Search
Answer ID 12067   |   Last Review Date 01/28/2022

What articles are returned when drilling/faceting into the product tree when search results are returned?

Environment:

  • Oracle B2C Service
  • Sites with Knowledge Advanced only
  • Filtering by Product or Category during Search
Issue:
The Knowledge Advanced Product and Category filtering work differently from each other on search result filtering. 
 
Resolution:
When search results are returned the tags that can be used to filter the results are also returned.  These are often called facets.  Products and categories work differently as they filter.
 
Product filtering narrows the results.  Category filtering subsets the results.  This means when filtering into a product tree if the articles for the parent facets are the best answers, they will still be included in the answer set.
 
The reason for this is generally product trees are much much larger than category trees and they sometimes have many many branches and layers.  Most product trees are very large.  So if an article applies to all the products you do not want to tag all the other nodes.  That would slow the system and the authors and it becomes a lot to maintain.  You just want to tag the top node that the article applies to.  Then as the user filters the results, if the best answer for the question is still that article that applied to all products then it will still appear even though you want just articles that apply to one specific product version.

If an article only applies to a specific product version then you want to tag the article at that level.  And even if there are sub-products there, or they are added in the future, the article will still be seen when the user sub-selects under that product.

Category trees tend to be much flatter and have a wider classification without as much overlap. For this reason these filters are subset-ted.

The purpose of tagging helps the user find an article more quickly amidst a lot of different articles.  If the question is searching for installation information on a phone and its about some error message, you do not want to rely only on the key terms like 11mini and installation, that would be a simple keyword search and will bring back many many articles.  The user can add those terms to the question but the user should learn to add more specifics to the search about the error message or the cause of the message.   The question is the most important to getting to information quickly.   The knowledge advanced search engine considers many characteristics about the question and the articles that are returned.   The relationship of the words to each other in the articles, where and how often they appear in the articles, whether they are key words to the business (also considering synonyms), and how used the article is are some of the things that have the biggest impact on how the articles are scored and returned to the user.  With just keywords the articles returned become more random.  By making the search better than the tags can become a better aide to finding the right article quicker.