Use Relevance Tools

This is for:

Developer System Administrator
Warning
Legacy feature

This article pertains to achieving relevance with the Coveo Hive framework. Coveo Hive is now in maintenance mode.

See Achieve relevance for guidance on leveraging Coveo Platform relevance features with the Coveo Atomic library.

Much of the search result relevance magic is handled behind the scenes by Coveo, from the moment you create machine learning models. However, when usage analytics reveals relevance issues, you can also intervene directly by setting rules of your own.

Articles in this section present the different options at your disposal when it comes to improving result relevance.

Using Coveo Query Pipeline Features

Whereas most of the Build Search phase of the Coveo implementation journey was spent working in Sitecore, your focus should shift to the Coveo Platform when it comes to improving relevance. query pipelines are rules that can be specified in the Coveo Administration Console to alter queries before they’re sent to the index. In particular, filtering and boosting rules can be configured in a query pipeline.

You can associate conditions to your Query Pipelines or associate conditions to your Query Pipelines to apply filtering and boosting expressions dynamically.

Using the Rule Set Editor

Enterprise edition

The Rule Set Editor is a Sitecore wizard which lets you create rules by specifying conditions (which can be combined with an action). Coveo for Sitecore integrates into this feature by adding its set of conditions and actions for result filtering and boosting purposes.

Using JavaScript in Custom Coveo for Sitecore Renderings

Although not as convenient and simple as working with the Sitecore Rule Set Editor, you can create custom Coveo for Sitecore renderings with your own JavaScript code to apply filtering and boosting expressions.

Automatic Boosting Based on the Current Sitecore Visitor Pattern

Enterprise edition

Important

Automatic boosting using Sitecore visitor patterns is no longer recommended. Coveo Machine Learning does a much better job without all the complex interventions.

Automatic boosting based on the current Sitecore visitor pattern requires a Sitecore instance with xDB installed. This method also requires a significant amount of item data entry and configuration in Sitecore.

This feature is available through the Rule Set Editor in the data source of the following Coveo for Sitecore renderings:

Tip
Leading practice

Although the Coveo Search Interface component view file automatically instantiates both a Query Filter and a Query Ranking component, you should insert standalone instances of these components in your search interface(s). This eases the configuration and reusability of these components.