Boost Your Search Experience with Yext's Algorithm Update

Table of Contents

1. Introduction to the Nebula Algorithm Update
2. The Vertical Ranking Model
   1. How the Model Works
   2. Configuring Vertical Ranking
      • Thresholds
      • Biases
      • Vertical Intents
3. Functions and Query Rules
   1. What Are Functions?
   2. How to Use Functions in Query Rules
   3. Examples of Function Usage
4. Testing and Debugging Functions
   1. Using the Dino Test Framework
   2. Locally Testing Functions
5. Conclusion
6. FAQ

The Nebula Algorithm Update: Improving Search with Yext's Latest Release

Welcome to the exclusive webinar on upgrading your search with Yext's latest algorithm update, the Nebula Algorithm. In this webinar, we'll be diving into the various features and enhancements included in this update and how they can improve your search experience.

1. Introduction to the Nebula Algorithm Update

At Yext, we release major updates to our search algorithms several times a year to ensure we're continuously improving search experiences on the web. The Nebula Algorithm update is part of our Spring 22 release and includes significant improvements to enhance search relevance and performance.

2. The Vertical Ranking Model

The Vertical Ranking Model is a key feature of the Nebula Algorithm update. It focuses on determining the most Relevant order of vertical results for a given query. By using machine learning, the algorithm assigns a relevance score to each vertical, helping to rank and prioritize verticals in search results.

2.1. How the Model Works

Every vertical in an answers search now receives a relevance score between 0 and 1. This score is Based on various factors, including the query's semantic similarity to the vertical, filters applied on the vertical, and the top result returned in each vertical.

2.2. Configuring Vertical Ranking

To fine-tune the vertical ranking for your specific queries, Yext provides three new configuration options: thresholds, biases, and vertical intents.

Thresholds

Thresholds allow You to set minimum score requirements for each vertical. If a vertical's score does not meet the threshold, it won't be returned in the search results. You can use thresholds to ensure verticals only appear when there is a strong match with the query.

Biases

Biases enable you to add an adjustment directly to a vertical's score, affecting its final ranking. Positive biases can be used to prioritize specific verticals, such as promotional verticals or official content. Negative biases can be applied to filter out irrelevant or wordy content.

Vertical Intents

Vertical intents are phrases associated with each vertical, factor into the vertical's semantic score. By specifying certain phrases or brand-specific terms, you can ensure that a particular query will prioritize and return the intended vertical first.

3. Functions and Query Rules

Another powerful feature introduced in the Nebula Algorithm update is the integration of functions and query rules. Functions allow you to write custom code that can be executed in response to specific events, such as a query rule matching a particular criteria.

3.1. What Are Functions?

Functions in Yext are serverless functions that run on the Yext platform in response to events. They provide a way to customize search behavior and fetch data from external sources, making it possible to extend the capabilities of answers beyond the Yext Knowledge Graph.

3.2. How to Use Functions in Query Rules

In query rules, functions can be added as an action in the "then" part of the rule. When a query rule matching specific criteria is triggered, the associated function runs and returns results, which can be displayed in the search interface.

3.3. Examples of Function Usage

Functions in query rules open up a wide range of possibilities. You can use functions to fetch data from external APIs, integrate with private systems, or personalize search results based on user-specific information. For instance, you can fetch reservation data from an external service, retrieve flight information from an airline's database, or display real-time stock prices obtained from an open API.

4. Testing and Debugging Functions

To ensure the proper functionality of functions, it is crucial to test and debug them. Yext provides tools to facilitate this process.

4.1. Using the Dino Test Framework

The Dino Test framework is specifically designed to test Yext functions. It allows you to write and run test cases, ensuring that your functions behave as expected. By logging the results, you can identify and fix any issues before deploying the functions.

4.2. Locally Testing Functions

To test functions locally, you can leverage the Dino Test framework. By simulating inputs and reviewing the outputs, you can validate the behavior and functionality of the functions before deploying them on the Yext platform.

5. Conclusion

The Nebula Algorithm update, along with the integration of functions and query rules, brings advanced customization and improved relevance to your search experience. With the flexibility to configure vertical ranking, customize search behavior with functions, and test functions efficiently, you can enhance your search capabilities and better meet the needs of your users.

6. FAQ

Q: Can I still use base score to order verticals?\ A: While base scores can still be used, it is recommended to leverage the new tools like biases and thresholds for better flexibility and control over vertical ranking.

Q: Can functions be sensitive to authentication or authorization?\ A: Yes, functions can be designed to incorporate authentication and authorization mechanisms to ensure secure access and display of relevant information for authorized users.

Q: Can a function be applied to only a separate instance of the X search box?\ A: Yes, it is possible to apply functions selectively based on criteria like authentication or specific page context. This allows for personalized search experiences tailored to different users or scenarios.

Q: Can functions be tested before deployment?\ A: Yes, functions can be tested using the Dino Test framework, allowing you to simulate inputs and verify outputs. This ensures the proper functionality of functions before they are deployed.

Q: Can I use phrases for vertical intents?\ A: Yes, phrases can be utilized as vertical intents. Vertical intents leverage semantic similarity to associate specific phrases or brand-specific terms with a particular vertical, influencing the vertical's ranking for relevant queries.