Rules of Thumb for the Search
The following rules of thumb illustrate how content is found in Serviceware Knowledge and how relevant a content is with regard to a certain search term.
-
Content is found if the search term occurs in at least one of the fields title, content, tags or categories.
Example: Text A contains the search term X in the title and text B contains the search term in the title and in the content. If a user searched for X text A as well as text B are found.
-
Depending on how many fields the search term occurs in, the content is more relevant in respect to the search query. The relevance of a search result controls its position in the content list.
Example: Text A contains search term X in the title, in the content and as a tag. Text B contains search term X only in the content. Text A is more relevant in respect to search term X and is therefore displayed higher in the content list than text B.
-
The more often a term exists in the content of the Serviceware Knowledge system, the less relevant each individual content containing the search term is with respect to the search query.
Example: The own company name (e.g. Telmo) exists in many content items in the Serviceware Knowledge content of the company. If a user searches for the company name in combination with another search term (e.g. Telmo address), the relevance of a certain content with respect to the search query is calculated as the sum of the relevances of the two search terms.
This means for text A the relevance with respect to the search term (here: Telmo address) is calculated by adding the relevance of the first word of the search term (here: the company name Telmo) to the relevance of the second word of the search term (here: address). Since the company name occurs frequently in the Serviceware Knowledge content, the occurrence of the second word (here: address) of the search term has more weight in the calculation of the relevance of the whole search term.
A text B where the search term “address” occurs five times whereas the search term “Telmo” only occurs three times will be ranked higher in the content list than a text C which only contains the search term “address” three times and the search term “Telmo” five times.
-
The relevance of each content with respect to a search term depends on the length of the respective content in comparison to the average length of all the content items in the Serviceware Knowledge system.
Example: Text A has 300 words and contains the search term once. Text B has 100 words and also contains the search term once. Thus, text B (the shorter content) is more relevant with respect to the search term than text A (the longer content), since the search term occurs more often in comparison to the content length of text B. Therefore, text B is ranked higher in the content list.