Back
Module Assessment | Yext Hitchhikers Platform
Background
Now that you’ve set up your site, it’s time to review the search quality and make some refinements. Let’s run some searches and see what changes we can make to optimize the experience. You’re looking at top search terms and want the results to look good (maybe you pulled the top search terms from your site before you implemented Yext Search or maybe your Search experience has been live for a couple months and you pulled the top search terms for it).
Your Challenge
Navigate to Search > Brand Search > Home. In Test Search, run the following top-searched queries. As you run the searches, take note whether the results returned are what you’d expect.
- order online
- locations that accept reservations
- upcoming events
- restaurants near me
- phone number for Miami location
- jobs
Search 5-10 other queries as well.
Navigate to the Search Logs screen. Open each search log in a new tab to review the queries. For each query:
- Mentally note whether the results look good or need review. For example, if there are no results or the results are not relevant to the query, you’ll want to make adjustments to improve the results. In those cases, what do you think is the issue? Is it an issue with the Content, configuration or algorithm?
Now that you’ve reviewed the queries, let’s start improving! The query “jobs“ should be returning jobs! Hiring can be hard and no results means you just lost a lead!
- Optimization: Write up an email draft to your HR department to request content for jobs so you can add it to your experience. We will work on this more in the next challenge!
The query “locations accepting reservations“ should be returning the restaurants vertical and ideally only those restaurants that are accepting reservations, but it’s not doing either. You can try ”location accepting reservations“ which at least returns restaurants, but not filtered to those accepting reservations like the search calls for.
Optimization: This seems like a two-fold problem:
- You need to tell the algorithm to treat “locations” the same way it would treat “restaurants”. Add “locations” to the existing synonym set that includes “restaurant” and “location”. Run the query again in Test Search to make sure it updates (could be a short delay). You should see “Restaurant” appear as an NLP filter pill.
- You remember that you created a field called Restaurant Features and have this data filled out already with “Accepting Reservations” as a feature. You just need to add it to the searchable fields. Add Restaurant Features (
c_restaurantFeatures) as an NLP filter on the Restaurants vertical. Run the query again in Test Search to make sure it updates (could be a short delay). You should see “Accepting Reservations” appear as an NLP filter pill.
The query “phone number for Miami location” surfaces the Miami location, but the entity preview doesn’t have the restaurant’s phone number. You may decide to display this on the frontend, but you also want to showcase the phone number to directly answer the user query.
- Optimization: Add a direct answer for the phone number field on the Restaurants vertical.
- To do this in the UI, navigate to the Verticals screen and select the
Restaurantsvertical from the dropdown. Find the Direct Answers section. Next to Eligible Fields click + Add / Update Fields and select themainPhonefield. To do this in the JSON editor, navigate to the Edit as JSON screen and find the
restaurantsvertical. Add an object fordirectAnswerslike so:"restaurants": { "directAnswers": { "fieldValues": { "eligibleFields": [ "mainPhone" ], "predictionMode": "CLASSIFIER" } }, }
Make any other changes to optimize the experience based on the queries you ran. Which of the other ones need improving and how can you improve them? Post in the Community and tell us what else you did!
Weekly Challenge Streak: <% util.streakWeekly %>
