Overview

This article outlines the process for setting up and configuring HawkSearch recommendations. The Recommendation functionality enhances your platform by providing personalized product suggestions, potentially increasing conversion rates, average order value, and customer satisfaction. By leveraging user behavior data, it improves product discovery and offers valuable insights for inventory and marketing decisions.

Configuration Steps

The core of the recommendation system relies on the HawkSearch recommendation API. This process involves setting up the API connection in your Optimizely dashboard, ensuring correct event tracking, and configuring the recommendations in the HawkSearch workbench. The following steps will guide you through this integration process.

Optimizely Dashboard Configuration

1. Log into your Optimizely dashboard
2. Navigate to: Administration > System > Settings > Search > Xngage HawkSearch section
3. Locate the Recommendation API URL field
4. Add the appropriate URLs according to the environment. For more details regarding Recommendation API, please refer to this article

5. Confirm that your event tracking is set up correctly, please refer to this article for more details.

HawkSearch Workbench Configuration

1. To configure your recommendation widgets, please navigate to the Recommendation administration in HawkSearch workbench
2. For more details on configuring recommendation through HawkSearch workbench, please refer to this article.

Verifying Recommendation Configuration

After setting up your recommendations, it's important to verify that they are working correctly. For details about different recommendation strategies, please refer to this article.

Testing with Featured Item Strategy

Testing with the Featured Items strategy is the most basic way to verify if recommendations are working:

1. Configure the Feature Items strategy:
   1. Log in to the HawkSearch workbench
   2. Set up a new recommendation widget using the Feature Items strategy
2. Prepare the API request
   1. Use an API testing tool (e.g. Postman)
   2. Set up a POST request to the HawkSearch recommendation API endpoint to get the recommendation widget you set up
3. Send the request to the API endpoint
4. Examine the API response:
   1. Verify that the response status is 200
   2. Ensure the response contains a list of featured items from your index
   3. Verify that the returned items align with your widget configuration (e.g., match condition, proper sorting)
   4. Look for any error messages or unexpected data in the response

Testing with Recently Viewed Strategy

The Recently Viewed strategy relies on user behavior tracking and requires a sequence of events before testing the recommendations:

1. Configure the Feature Items strategy:
   1. Log in to the HawkSearch workbench
   2. Set up a new recommendation widget using the Recently Viewedstrategy
2. Simulate user interactions by sending tracking events
   1. Send a search event: make a POST request with search event type
   2. Send a click event:
      1. choose an item from the search results to be tested
      2. make a POST request with click event type
   3. Send a page load event: make a POST request with page load event type to simulate viewing the product
3. Prepare the API request
   1. Use an API testing tool (e.g. Postman)
   2. Set up a POST request to the HawkSearch recommendation API endpoint to get the recommendation widget you set up
4. Send the request to the API endpoint
5. Examine the API response:
   1. Verify that the response status is 200
   2. Ensure the response contains the item(s) from your simulated user journey
   3. Verify that the returned items align with your widget configuration (e.g., match condition, proper sorting)
   4. Look for any error messages or unexpected data in the response