Get search results

This page describes basic querying with search, including text query searches, browse searches, pagination, optimization, and personalized results.

Never cache personalized results from an end user, and never return personalized results to a different end user.

Relevance in Vertex AI Search for retail

Relevance is different from matching. Matching is finding candidate products in the index that correspond to the user's search query. Relevance, on the other hand, is a degree of matching. It involves ranking those products in a way that best satisfies the user's intent and needs. Vertex AI Search for retail uses techniques to improve matching and relevance for a better overall search experience for users that meets retailers' need.

Non-default search

Vertex AI Search for retail defaults to relevance in search. However, search is tuned to also surface marginally relevant items higher up in the search results to provide retailers more targeted search results to increase the conversion potential of user events.

These added search features optimize the default relevance algorithm, providing non-default features tuned to each retailer's utilization profile, including:

• Options to sort results such as by price, top-rated, or best-selling items. Non-default search is customized to return specific search results with the correct product variants. For example, to show the cheapest or most expensive variant when sorting by price.

• Auto-application of a tighter relevance filter for non-default user events.

• Enabling adjustable strength boost/bury controls to fine-tune search results.

Text search and browse search

Search provides both text query search and browse search capabilities.

In the text query search use case, a shopper might enter a text-based query on your site. Search returns a search response containing products that fit within the parameters of controls you have set up, sorted by relevance and revenue maximization.

In the browse use case, a shopper might go to your site menu and navigate to a specific product category. Search automatically chooses the most revenue-maximizing sort order by learning from user behavior and trends. Browse results can be further refined by the controls you have set up.

Both text search and browse search requests use the servingConfigs.search method.

Text query searches

When a user enters a text query to search on your site, search sorts potential search results based on relevance, popularity, buyability, and personalization.

Search considers a servingConfigs.search request as a text-based search request if it has a non-empty query field.

When uploading user event, send text query search events generated by search as search user events. If the event has a non-empty userEvent.searchQuery field and an empty userEvent.pageCategories field, search considers it a text-based search event.

Browse searches

Browse uses site navigation to produce search results of equal relevance sorted by best-selling items. Search leverages AI to optimize how browse results are sorted by considering popularity, buyability, and personalization.

Serving requests and events with an empty query field and non-empty page_categories fields are considered browse requests and user events. The page_categories field specifies the browse page itself. When the servingConfigs.search method sends a request, search considers it a browse search request if the query field is empty. When that is the case, the results are based on the filter and pageCategories fields, as well as further optimization and personalization if available.

For correct browse search results, consider the following:

• The pageCategories and filter values in your search requests must exactly match the pageCategories and filter values in your uploaded user events. filter values in your uploaded user events. If the parameters in the requests to not match parameters in the events, the click-based reranking model learned from the events won't perform very well for this specific query and will negatively impact result quality.

• Vertex AI Search for retail categorizes a search user event as a browse-based event if it has an empty userEvent.searchQuery field and a non-empty userEvent.pageCategories field. This is why you need to make sure you send browse events generated by Vertex AI Search for retail as search user events when uploading them.

• A browse filter must be valid and indicate for each browse query which products belong to the product category the user is browsing in. To achieve this setup, the category ID needs to be defined in both the filter and the page_category fields. Take, for example, if you want the search results to be targeted to Home & Garden. First off, having only one filter value of "in-stock" does not refine the search enough to provide optimal browse results. To provide search results restricted to Home & Garden, you not only need to set "Home \& Garden" in the page category, but also in the filter with categories: ANY("Home & Garden"). If you neglect to define the category in the filter part, browse results could likely be too broad and include many non-Home & Garden items.

You can set the category in both as in this example:

page_category: "Home & Garden"
filter: "(availability: ANY("IN_STOCK")) AND (categories: ANY("Home & Garden"))"

Otherwise, because browse works with empty string queries, the products in a given category won't automatically be displayed in the results for a given page category. In other words, if the end user browses without specifying anything using a text query, the search return may not yield effectively targeted results, even if the user navigates to a category page.

Use search requests

Use search requests to get results for both text searches and browse searches. To make a search request, use the servingConfigs.search method.

All search requests require placement, which identifies the full resource name of the serving config that will be used. The serving config determines which settings and associated controls affect the search results.

Text query search requests require a non-empty query field.

Browse search requests require a non-empty pageCategories field.

import com.google.cloud.retail.v2.SearchRequest;
import com.google.cloud.retail.v2.SearchResponse;
import com.google.cloud.retail.v2.SearchServiceClient;

public static void searchProducts(String query) throws IOException, InterruptedException {
  SearchRequest searchRequest = SearchRequest.newBuilder()
      .setPlacement(DEFAULT_SEARCH_PLACEMENT_NAME)
      .setBranch(DEFAULT_BRANCH_NAME)
      .setVisitorId(VISITOR_ID)
      .setQuery(query)
      .build();

  try (SearchServiceClient searchClient = SearchServiceClient.create()) {
    SearchResponse response = searchClient.search(searchRequest).getPage().getResponse();
    System.out.println("Search response: " + searchResponse);
  }
}

By default, a reasonable number of results ordered by relevance is returned.

To get product attributes returned with the search response, make sure to provide attribute values when you import your catalog data. Product has predefined system attributes such as brand, color, and size that you can provide values for. You can also include custom attributes that you define with Product.attributes.

Querying tutorial

This tutorial shows you how to send a text-based search query to the Vertex AI Search for retail service and analyze the response.

To follow step-by-step guidance for this task directly in the Cloud Shell Editor, click Guide me:

Guide me

Pagination

Use pagination to decrease lookup time and the size of responses being sent.

Pagination tutorial

This tutorial shows how to control pagination in a text-based search request. When a shopper looks for products in a shop, they can improve their navigation through the search results. For example, they can limit the number of items in the search response by using the page size feature or jump to their preferred page by using the offset feature.

To follow step-by-step guidance for this task directly in the Cloud Shell Editor, click Guide me:

Guide me

Paginate

To jump from one page to another, use either page_token or offset, according to your use case.

To jump to the next page, you could use page_token. For example, suppose you send the following SearchRequest.

{
  placement: 'projects/PROJECT_NUMBER/locations/global/catalogs/default_catalog/placements/default_search'
  visitor_id: 'VISITOR_ID'
  query: 'shoes'
  page_size: 5
}

From SearchResponse, you can get the resulting products with top 5 relevance, along with a next_page_token.

{
  results: [
    products{...}, products{...}, products{...}, products{...}, products{...}
  ]
  next_page_token: "wY4ETNkBDOlVjZ0YWLzUmM40SMhVjMtADMwATL5UGN5MGZlVDJaIQ5LaYsQUw9fC6lIwgE1EgC"
  total_size: 100
  search_token: "NtQKDAiXt4_3BRDCg_jnARABGiQ1ZWRjOTRlOC0wMDAwLTI1YTEtODJlMy1mNGY1ZTgwZDUxOGM"
}

To get the result products with the next 5 relevance (6th to 10th), you would set page_token using the same placement, visitor_id, and query as next_page_token from the previous SearchResponse.

import com.google.cloud.retail.v2.SearchRequest;
import com.google.cloud.retail.v2.SearchResponse;
import com.google.cloud.retail.v2.SearchServiceClient;

public static void searchProducts_withNextPageToken(String query, int pageSize)
    throws IOException, InterruptedException {
  try (SearchServiceClient searchClient = SearchServiceClient.create()) {
    SearchRequest firstRequest = SearchRequest.newBuilder()
        .setPlacement(DEFAULT_SEARCH_PLACEMENT_NAME)
        .setBranch(DEFAULT_BRANCH_NAME)
        .setVisitorId(VISITOR_ID)
        .setQuery(query)
        .setPageSize(pageSize)
        .build();

    SearchResponse firstResponse = searchClient.search(firstRequest).getPage()
        .getResponse();
    System.out.println("First search response: " + firstResponse);

    SearchRequest secondRequest = SearchRequest.newBuilder()
        .setPlacement(DEFAULT_SEARCH_PLACEMENT_NAME)
        .setBranch(DEFAULT_BRANCH_NAME)
        .setVisitorId(VISITOR_ID)
        .setQuery(query)
        .setPageSize(pageSize)
        .setPageToken(firstResponse.getNextPageToken())
        .build();

    SearchResponse secondResponse = searchClient.search(secondRequest).getPage()
        .getResponse();
    System.out.println("Second search response: " + secondResponse);
  }
}

In this example, SearchRequest looks like this:

{
  placement: 'projects/PROJECT_NUMBER/locations/global/catalogs/default_catalog/placements/default_search'
  visitor_id: 'VISITOR_ID'
  query: 'shoes'
  page_size: 5
  page_token: "wY4ETNkBDOlVjZ0YWLzUmM40SMhVjMtADMwATL5UGN5MGZlVDJaIQ5LaYsQUw9fC6lIwgE1EgC"
}

In some other cases, instead of navigating from page to page or getting results with top relevance, you could directly jump to a particular position with offset.

import com.google.cloud.retail.v2.SearchRequest;
import com.google.cloud.retail.v2.SearchResponse;
import com.google.cloud.retail.v2.SearchServiceClient;

public static void searchProducts_withOffset(String query, int pageSize,
    int offset) throws IOException, InterruptedException {
  SearchRequest searchRequest = SearchRequest.newBuilder()
      .setPlacement(DEFAULT_SEARCH_PLACEMENT_NAME)
      .setBranch(DEFAULT_BRANCH_NAME)
      .setVisitorId(VISITOR_ID)
      .setQuery(query)
      .setPageSize(pageSize)
      .setOffset(offset)
      .build();

  try (SearchServiceClient searchClient = SearchServiceClient.create()) {
    SearchResponse response = searchClient.search(searchRequest).getPage().getResponse();
    System.out.println("Search response: " + searchResponse);
  }
}

For example, if you want the tenth page of the results when the page size is 5, then you could set the offset to be 45, which is calculated by (10 - 1) * 5.

{
  placement: 'projects/PROJECT_NUMBER/locations/global/catalogs/default_catalog/placements/default_search'
  visitor_id: 'VISITOR_ID'
  query: 'shoes'
  page_size: 5
  offset: 45
}

Search performance tiers

Search offers several search performance tiers that increasingly improve your results. For example, for text query search use cases, results might be based solely on relevance. As you unlock more advanced performance tiers, search can return results that are based on relevance, popularity, revenue optimization, and personalization.

Search automatically unlocks higher performance tiers when you upload catalog and user event data that meets the minimum requirements of each tier.

For more information about using this page to view data quality and search performance tiers, see Unlock search performance tiers. The Data quality page on the Search for Retail console provides an assessment of which requirements you have met for each tier.

Evaluate text search and browse results

Before you update your website code to request text search or browse search results, you can preview the results to confirm that your serving config is working as you expect.

For more information about serving configs, see About serving configs.

You can preview serving config results either from the Evaluate page, or by going to a serving config's Details page in the console and clicking its Evaluate tab. The following steps show you how to preview from the Evaluate page.

To preview the results returned by your serving configuration:

To see the Details page for the serving config you're previewing, click View serving config under the Select serving config field.