Site search
Site search is the feature that lets a visitor find content inside a website, web portal or intranet from the words they type. It relies on an index, an organized copy of content, its fields and its metadata, which the engine queries to rank results by relevance.
Definition and how it works
The quality of site search depends on three settings more than on the engine itself. Indexed fields carry different weights, so a word in the title counts more than a word in the footer. Vocabulary covers synonyms, acronyms, typos and the business terms visitors use and writers do not. Freshness means that content that is published or withdrawn must appear in or disappear from the index without delay.
Behind a sign-in, a security requirement comes on top. Results and the snippets shown under each title must both respect the signed-in person’s permissions. A confidential document missing from navigation that still surfaces in search, with three lines of its content, is a leak.
Related terms: Tags · Classification · Intranet · Content Type · Web portal · Content management system
Resources that cover this term
Site search vs web search: the differences
Site search queries a defined scope, an organization’s content, with its access rights. Web search queries the public web using a third-party engine’s ranking. A filter narrows a list already on screen without querying a wider index. A good experience can combine search with facets, letting the visitor type a word and then narrow by content type, date or category.
| Concept | What separates it |
|---|---|
| Web search | It covers the public web, ranked by a third-party engine. The organization controls neither the index nor the permissions. |
| Classification | It applies metadata (tags, categories, keywords, properties) that then serve as facets and relevance signals. |
| Tags | Terms attached to content as metadata to classify or find it. In Jahia CMS, they are shared across languages. |
| Federated search | It queries several sources (website, documentation, ticketing tool) and returns one result list, each source keeping its own permissions. |
Site search in practice
The best source for evaluating site search is its own query log. Zero-result queries reveal visitor vocabulary the content does not use, or content that is missing. Queries immediately followed by another search signal useless results. Those two lists, reviewed monthly, are worth more than tuning done once at launch. The article Internal site search: 4 best practices for better results covers ways to improve them.
For a pre-purchase demo, prepare twenty or so real queries from that log or from support, including a few acronyms, a typo and a query about restricted content. Running them with two different profiles shows within minutes whether relevance and permissions hold up.
Frequently asked questions
Can site search expose content a user is not allowed to see?
It can when the index ignores permissions, especially with search added as an afterthought. In that case, the engine indexes everything, then the interface hides some results imperfectly or still shows their snippets. Safe search filters results according to the signed-in person’s permissions at query time. Test it with a restricted account, never with an administrator account.
What is site search for when Google already brings visitors to the site?
It becomes essential as soon as the site is large or part of the content sits behind a sign-in. Google cannot index what its crawler cannot reach, such as a customer area or an intranet behind a sign-in. Nor does it know each user’s permissions. On a public site, site search is also a data source, since the queries people type reveal needs the navigation does not cover.
What is the difference between built-in CMS search and a dedicated engine?
Built-in content management system search covers the CMS’s own content first and how far its relevance can be tuned varies by product. A dedicated engine, for example one built on Elasticsearch or OpenSearch, adds facets, typo tolerance, synonyms and indexing of external sources such as documents or other applications. The deciding factor is scope: one site or several sources, with or without access rights.