A site's owner and edit group members can add a search bar anywhere on a site or page using the Search card. Using this search bar, people can use keywords and phrases to explore the site's content. You can also enable global navigation to add a search bar to the top of the site next to the user profile and notifications buttons.
What shows up in search results?
Content added to the site's content catalog appears in your site's search results. For example, if you created an open data site to share local data, a community member can enter a search query for parks on your site's search bar and any item related to local community resources (such as web maps of playgrounds, schools, and pools) will appear as search results.
When searching for content, users can search and filter by location, and choose a grid or a list view of search results. For public sites, users can export public items from the site catalog to a CSV file using the more actions 
button.
You can add groups of content to the site's catalog through the catalog configuration. Content that belongs to designated open data groups will also appear as search results for any search query entered at hub.arcgis.com. For more information, see Add and manage content and Make public content more visible.
Note:
Staff and community members who have access to items that have been shared internally can view these private items in search results when signed in with their account.
Add a search bar
Add a search bar to a site or page layout by configuring a search card.
To configure a search card, complete the following steps:
- Open a site or page in edit mode.
- On the Customize panel, select Layout.
- Drag a Search card onto a row in the layout.
- Select whether to show the browse button instead of a search bar.
- Choose a size for Search bar width.
- Provide default text to appear in the Search placeholder text field. For example, you could use Search by Keyword or Find Data.
- Select Save layout.
How searching works
When executing a search, content items are compared to a query and ranked by relevance. Relevance is calculated by syntactic matches between a search query and content item metadata fields including title, description, and more.
Boosting, or relative importance among fields, is a complex process and happens at several different stages during generation of search results. It only occurs when there is a match on a particular field. The fields that are boosted the most are "title" and "description" followed by "categories" and "tags."
After search results have been calculated, filters and their list of values are returned if there are two or more values. Each filter lists values by aggregating metadata from all returned search results. Filter values are ranked in descending order of usage frequency.