Now that you have set up your Search configuration, it's time to refine the results. In general, start with fewer configurations (searchable fields, vertical ranking adjustments, synonyms, and query rules) so you can get a sense for the base performance of your experience and then make adjustments from there.
Once your experience is live, you'll want to monitor what users are searching for and continually optimize. Things change — customers change, brands release new products, larger market trends shift. Your Search experience, like any other part of the Yext platform, will be in a state of continuous optimization.
In this module, we'll cover what continuous optimization looks like for Search: debugging unexpected results, refining results with experience training and the search merchandiser, and monitoring search terms and analytics to understand what users are searching for and where the gaps are.
Understanding Customer Intent
The first step to making Search optimizations is understanding customer intent. Search Analytics shows valuable insights — how complete your search results are, how many clicks you're getting, what people are exactly searching for — so you can better understand what changes to make to improve the consumer experience.
As you gain insight from people's searches and clicks, you can get a sense for their true intent, then make educated decisions about search and about your business strategy as a whole. As you improve the answers people are seeing, they'll come to rely more heavily on search, giving you even more information about what they need.
In many cases, Yext encourages going live as soon as possible to start collecting this valuable data, even if you haven't established a complete CMS. This allows you to better inform how you structure Knowledge Graph. If a brand already has search on their site, past search logs can also help.
Types of Optimizations
In Search, there are two categories of optimizations:
UI Improvements — Adjusting the look and feel of the search results to drive more conversions or engagement. Examples include changing CTA labels or URLs to maximize click-through rate, exposing or hiding content on result cards, and changing colors or fonts. Since these are frontend optimizations, they're covered in more depth in the frontend tracks.
Search Quality Improvements — Search quality is affected by Knowledge Graph, the Search configuration, and the Search algorithms. When reviewing search quality issues, determine which of these three is driving the issue.
Knowledge Graph Issues
Knowledge Graph issues mean your CMS simply doesn't have the content being searched for. For example, if you're seeing many searches for "marketing jobs" with no results, some possible causes are:
- No entity type: You don't have a Jobs entity type enabled. Add it to your Knowledge Graph and add a new vertical in your Search experience.
- No relevant field: You don't have a Department field on the Job entity type. Add the field and make it searchable.
- No matching data: You have a Department field, but no open Jobs in marketing. If there just aren't marketing jobs available, you could update the No Results component or add an FAQ indicating that marketing isn't currently hiring.
Configuration Issues
If you have the content in your CMS, it's almost certainly a configuration or algorithm issue. Configuration issues mean you can make an update in the backend to fix results. Be careful that updates to improve one search term don't negatively impact another.
Note: Before making configuration changes, make sure the production configuration label is set to an explicit version (not "Latest"). Make sure the staging label is set to "Latest" so you can test your latest changes in staging.
Configuration improvements for the same "marketing jobs" example:
- No vertical: Add a Jobs vertical to your Search experience.
- Missing searchable fields: Set the Department field as searchable with the appropriate algorithm.
- Add a synonym: If you don't have open marketing jobs but do have SEO specialist roles, add a synonym from marketing → SEO specialist.
Algorithm Issues
If a search quality issue is not caused by the Knowledge Graph or configuration, it may be an algorithm issue — a bug or core component causing unexpected results. Raise any such issues to the Yext Search team.