Search in Salesforce Knowledge
After completing this unit, you’ll be able to:
- Describe why admins optimize Knowledge search.
- Explain how search results are ordered by relevance.
- List the benefits of lemmatization and spell correction.
You’ve enabled Salesforce Knowledge and started building your knowledge empire. Now people need to find all this information. And quickly. Time is money, and the way to save both is by optimizing search.
This module will first offer you some Knowledge search basics as an appealing appetizer. Then, we serve up three courses: how to identify the best search options to enable, how to set up search customizations, and how to use analytics to make better decisions about your new search setup.
You’re going to save service reps some serious time, which means your customers will get more relevant articles more quickly. Happy reps and satisfied customers—all by making a few, quick changes to search. Bon appétit!
Salesforce search is powered by a search engine that brings users smart, relevant results automatically. You might be wondering how we determine which results to put at the top of the list. Here’s our secret sauce to find and rank the records users are looking for.
- Dash of search term frequency, order, and uniqueness
- Sprinkle of record activity
- Spoonful of access permissions
- Handful of other factors
Sounds delicious, doesn’t it?
We have another tool to give users more relevant results. It deftly handles searches with multiple terms by strategically employing AND and OR operators. First, we run an AND search for the search terms, which means a search for customer email returns results that contain both customer and email.
But what if the results of that query yield not quite enough? Not to worry. When there are too few results, we automatically switch to an OR search. Meaning, we return records that contain customer or email (not necessarily both). We feel this strikes a balance between providing too many and too few results. It’s the Goldilocks porridge strategy, you could say.
Knowledge searches are given some extra care. We average the number of instances of a term in a record so that longer documents aren't given undue priority over shorter ones. We also normalize the number of instances of a term in an attachment. Attachments with many instances of a search term aren't prioritized over more relevant records with fewer instances. Knowledge deserves some pampering, after all.
Search applies advanced features to Knowledge, such as spell correction, lemmatization, and synonym groups. With these features, search can present records that include variations on the user’s search term to widen the net of results. You may be wondering, for example, what lemmatization is. Let’s go around the room to make some introductions.
- With spell correction, users might see results that match a corrected spelling of the term. Say the user types Julai. Salesforce returns results for Julia because no records contain the word Julai. But, say a user types Julie. Salesforce returns the single record for Julie, but not records for Julia.
- Lemmatization identifies and returns variants of the search term. For example, a search for fry also returns results for frying and fried. With lemmatization, users get results even when they’re a little off on the exact phrasing.
- Synonym groups are words or phrases that are treated equally in searches. A search for one term in a synonym group returns results for all terms in the group. For example, a search for dessert returns results for all terms in the synonym group, which contains cake, pudding, pie, and cookies. But not cupcakes, because that trend is over. We’ll discuss synonym groups in more detail in the next unit.
|Spell Correction (Julie)||Lemmatization (Fry)||Synonym Group (Dessert)|
Want to learn more about how search works? Try out the first unit in the Search Solution Basics module, Choose the Right Search Solutions.