A content silo is an information-architecture group of pages covering one subject or user task, connected through useful internal links.
A silo is a navigation model
A content silo groups pages that serve one coherent subject or user task, then connects them through an understandable hub-and-leaf route. It is an information-architecture pattern, not a ranking switch and not simply a directory name. A useful silo helps a person enter at a broad page, choose a narrower question, and return to the wider context without relying on search.
For example, a technical site may have a hub for indexation control, supporting pages for robots directives, canonical URLs, sitemaps and status codes, and deeper examples under each topic. The pages are not interchangeable: the hub explains the decision area, while a glossary page defines a term and a guide shows an implementation. If every page repeats the same answer, a silo only hides duplication in a hierarchy.
Design the hub and leaves
Begin with real tasks and the current page inventory. Group pages only when their relationship can be stated in a sentence a visitor would recognise. Then decide what belongs at the hub, what deserves its own leaf page and what should remain a section on another page. Google’s helpful-content guidance supports people-first organisation; it does not prescribe a fixed number of levels or pages.
| Architecture element | Distinct job | Evidence it is needed |
|---|---|---|
| Hub page | Orient a visitor and compare subtopics | Several related questions need a shared starting point |
| Leaf page | Answer one narrow question completely | The topic has unique procedure, examples or constraints |
| Contextual link | Move between genuinely related tasks | The source sentence explains why the next page helps |
| Cross-silo link | Connect a secondary but useful relationship | It helps the current reader, not a sitewide metric |
A concrete planning example: a site has separate pages for “how to write robots.txt” and “what noindex means.” A hub can introduce the crawl-versus-index distinction and link to both. A page about HTTP 410 can cross-link to the index-control hub when it explains permanent removal, but it should not be placed there merely because every SEO term needs more links.
Link within and across topics
The primary route should run from a hub to its supporting pages and back. Use descriptive internal anchors that name the destination’s benefit. Cross-silo links are valid when the reader’s next question genuinely crosses a boundary: performance troubleshooting may link to rendering architecture, for instance. The test is local relevance, not whether a diagram looks perfectly compartmentalised.
- List current pages, their user task, canonical URL and existing inlinks.
- Draft candidate hubs around subjects that have multiple distinct supporting questions.
- Assign every proposed leaf a unique purpose and decide whether it is a page, a section or unnecessary.
- Add hub-to-leaf and leaf-to-hub links in rendered content; add cross-links only with an explicit reader benefit.
- Crawl from hub URLs and check that each indexable leaf has a short, direct route and no contradictory canonical or navigation path.
Test the architecture
A silo is working when a reader can move from an entry page to a specific answer without dead ends, and a crawl can discover the same route. Review orphan pages, duplicate titles or near-identical sections, and links that point through redirects. Do not create thin pages solely to complete a diagram. The documented ‘do not create’ decision is often more useful than a page with no unique answer.