How search works for articles depends on your use of search options, search terms, wildcards, and operators. Salesforce Knowledge search uses the same custom search algorithms that are available throughout Salesforce, which include mechanisms such as tokenization, lemmatization, and stopword lists, to return relevant search results.
|Available in: Salesforce Classic|
Salesforce Knowledge is available in Performance and Developer Editions and in Unlimited Edition with the Service Cloud.
Salesforce Knowledge is available for an additional cost in: Professional, Enterprise, and Unlimited Editions.
Many factors influence the order in which articles appear in the results list. Salesforce evaluates your search terms and your data to move more relevant matches higher in your list of results. Some of these factors include:
- When you don’t specify an operator in your article search, the search engine determines the best operator to use.
- Many searches use “AND” as the default operator. This means that when you search for multiple terms, all the terms must match to generate a result. Matching on all terms tends to produce search results that are more relevant than searches using the “OR” operator, where matches on any of the search query terms appear in the results.
- If the search engine doesn’t return any results that match all the terms, it looks for matches using the “OR” operator. With the “OR” operator, the search engine boosts documents that contain more terms from the search query, so that they appear higher in the results list.
- This algorithm calculates the frequency with which a term appears within each article. The algorithm then weighs them against each other to produce the initial set of search results.
- Articles that are frequently viewed or that are frequently attached to cases appear higher in the results. Article ownership and recent activity also boost an article in the results list.
- Proximity of Terms
- Articles that contain all the keywords in a search are ranked highest, followed by those with fewer keywords, followed by those with single keyword matches. Terms that are closer together in the matched document, with few or no intervening words, are ranked higher in the list.
- Exact Matches
- Matches on exact keywords are ranked higher than matches on synonyms or lemmatized terms.
- Title Field
- If any search terms match words in an article title, the article is boosted in the search results.
- Token Sequence
- If the search term is broken up into multiple tokens because it contains both letters and numbers, the system boosts results based on the same sequence of tokens. That way, exact matches are ranked higher than matches on the tokens with other tokens in between.