Google Cloud Search has several default expansions, interpretations, and
optimizations that affect search results. If you see unexpected search results,
refer to this guide before contacting Cloud Search support.

In this document, user queries are shown in brackets (like `[query]`)
and system interpretations are shown in quotes (like `"interpretation"`).

The query results in this document are illustrative examples of behavior and
provide insight into how expansions, interpretations, and optimizations
work; actual results might vary.

## Default expansions

Suppose a user searches for `[Joe's PDFs]`, but some results contain
highlighted words like `"documents"` instead of `"PDFs"`. This happens because
Cloud Search expands the query to include synonyms and word
stems, similar to Google Search. This expansion helps retrieve documents that
match the idea and intent of the query. Ranking algorithms then ensure the
best matches appear at the top.

When a user searches for `[Joe's PDFs]`, Cloud Search might
supply these additional words:

- For `[Joe's]`, it might match `"joe"` (stem expansion) and `"joes"` (synonym based on punctuation).
- For `[PDFs]`, it might match `"documents"` (synonym expansion) and `"pdf"` (stem expansion).

Default synonyms aren't always bidirectional. For example, a search for
`[phishing]` might match `"phish"`, but a search for `[phish]` might not
match `"phishing"`.

### Expansions for hyphenated and underscored words

Cloud Search treats hyphenated words and their non-hyphenated
equivalents differently, such as `[walk-in closet]` and `[walk in closet]`.
It also applies different optimizations for hyphenated and underscored words,
like `[walk-in]` and `[walk_in]`.

### Compensate for default expansions

Cloud Search doesn't guarantee any specific expansion. To ensure
bidirectional synonyms or domain-specific expansions, create your own
synonyms. For more information, see [Define
synonyms](https://developers.google.com/workspace/cloud-search/docs/guides/synonyms).

> [!NOTE]
> **Note:** Educate your users on using quotes for exact matching. For example, searching for `[Joe's "PDFs"]` returns documents that exactly match the word `"PDFs"`.

## Default interpretations

Cloud Search provides natural-language interpretation based on
your data source schema. For more information, see [Structure your schema for
optimal query interpretation](https://developers.google.com/workspace/cloud-search/docs/guides/query-interpretation).

### Disable natural-language interpretations

To disable natural-language interpretations for a specific query, set
[`QueryInterpretationOptions.disableNlInterpretation`](https://developers.google.com/workspace/cloud-search/docs/reference/rest/v1/query/search#queryinterpretationoptions)
to `true` in the search request.

## Default optimizations

Cloud Search provides these default optimizations:

- Spelling correction blending: For the query `[corpoate benefits]`,
  Cloud Search matches `"corpoate"` and the correct spelling,
  `"corporate"`.

  > [!NOTE]
  > **Note:** A query response might contain a suggested spelling in the [`SpellResult`](https://developers.google.com/workspace/cloud-search/docs/reference/rest/v1/query/search#spellresult) field. Although the spelling and synonym systems are separate, the synonym system might provide a correct spelling as a synonym. The `SpellResult` field only populates for actual spell corrections.

- Supplemental results: For queries with few or no results,
  Cloud Search uses a permissive set of related terms. For more
  information, see [Handle supplemental
  results](https://developers.google.com/workspace/cloud-search/docs/guides/query-guide#display_query_results).

## Normalize documents and queries

Normalizing standardizes words or phrases before or after a query. To ensure
consistent responses, consider these normalization steps:

- To normalize documents:

  1. Choose a canonical spelling for critical words in your repositories.
  2. Correct spellings in source documents or during indexing to match the canonical form.
- To normalize queries:

  1. Intercept user queries before sending them to Cloud Search.
  2. Rewrite words in queries to match the most common spelling in the indexed data source.
  3. Send the rewritten query to Cloud Search.

## Disable expansions, interpretations, and optimizations

To disable these features for a specific query, set
[`QueryInterpretationOptions.enableVerbatimMode`](https://developers.google.com/workspace/cloud-search/docs/reference/rest/v1/query/search#queryinterpretationoptions)
to `true` in the search request.