How to Build a Silo Structure: A Practical Guide

TL;DR — A silo structureA silo structure is a site organized by grouping related content into topic areas and concentrating internal links within each topic so a hub page and its supporting pages reinforce each other. Modern practice keeps the topical concentration and drops the old rule against cross-linking between topics. means grouping your pages by topic and linking them up so each topic reinforces itself: one main “hub” page links to all the detailed pages under it, and those pages link back. To build one, figure out which subtopics people actually search for, create a page for each, and link them together. You can put them in a matching folder — but the links are what matter, not the folder. And don’t be afraid to link out to related topics elsewhere on your site.

What “building a silo” actually means

A silo is just a group of pages about the same topic, wired together with internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them.. Picture a guide about “email marketing.” The hub is the broad overview page. The spokes are the detailed pages under it — “email subject lines,” “welcome sequences,” “list segmentation,” and so on. Building the silo means:

1. Deciding which subtopics get their own page (based on what people search for).
2. Linking the hub to every spoke.
3. Linking every spoke back to the hub, and to relevant sibling spokes.

That’s it. The whole thing is about grouping related content and linking it deliberately instead of leaving pages scattered with no connections.

The part everyone gets wrong

Two myths worth clearing up before you build anything:

• The folder doesn’t make the silo. You’ll see guides say “put your pages in /email-marketing/ and you’ve built a silo.” Not quite. Google reads your links, not your folder names. A tidy folder is nice for you and your CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. — but if the pages inside it don’t link to each other, there’s no silo. (My colleague in this cluster’s site architectureSite architecture is how a website's pages are organized, categorized, and interlinked. It controls how crawlers discover pages, how link equity flows, and how clearly search engines understand each page's topical context. Silo structure, hub and spoke, and topic clusters are the three common models. article makes the full case for why the link graph is the architecture.)
• You don’t have to wall off your topics. The old, strict version of siloing says “never link from one silo to another.” Ignore that. If a page in your email silo genuinely relates to a page in your ecommerce silo, link them. It helps readers and it doesn’t hurt your SEO.

How to know if a subtopic deserves its own page

Don’t invent pages to hit some “5 pages minimum” number you read somewhere. The test is simple: does the subtopic have its own search demand? If people search for “welcome email sequence” separately from “email marketing,” it’s a real spoke. If you’re padding to fill a quota, you’ll just end up with thin pagesThin content is web content that provides little or no value to users. Google's spam policies name it 'thin content with little or no added value' — and it's about value per page, not word count. competing with each other.

The quick version of building one

• Research the topic and list the subtopics people actually search for.
• Make a hub page for the broad topic and a page for each real subtopic.
• Link the hub to every subtopic page, ideally right where you mention that subtopic.
• Link each subtopic page back to the hub and to closely related sibling pages.
• Link out to related topics elsewhere on the site when it’s genuinely useful.

Want the full build-and-audit process — how to draw silo boundaries with keyword data, a link-matrix template, cross-linking rules of thumb, and the tools to verify it worked? Switch to the Advanced tab.

TL;DR — This is the how, not the whether. Assume topical concentration plus sensible cross-linking is the goal (the model debate is settled in this cluster’s site architectureSite architecture is how a website's pages are organized, categorized, and interlinked. It controls how crawlers discover pages, how link equity flows, and how clearly search engines understand each page's topical context. Silo structure, hub and spoke, and topic clusters are the three common models. article). Build order: (1) draw silo boundaries from real keyword dataSearch volume, difficulty, and related queries for what people type into search engines. — distinct search demand per subtopic, not a page quota; (2) mirror the silo in folders for CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. sanity if you like, but remember the link graph is the architecture, not the path; (3) wire the internal linkAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. matrix — hub to every spoke, spoke back to hub, spoke to relevant siblings; (4) cross-link to other topics when a reader would actually follow the link, and never nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing. a cross-silo link; (5) avoid the classic mistakes — over-siloing, orphaning cross-topic content, boundaries drawn around your org chart instead of search intent; (6) audit with a crawl to confirm the links you planned actually exist.

First, the debate this article is not having

If you came here asking “is siloing dead?” or “silo vs. topic cluster — which is right?”, that’s answered in full elsewhere in this cluster. This site’s site architecture article runs the whole silo vs. hub-and-spoke vs. topic-cluster comparison and lands where I land: keep the topical concentration silos were reaching for, drop the strict “never link between silos” rule, and stop shopping for the “correct” brand name. Go read that if you want the why.

One naming note, since “silo” gets used loosely across the industry: the term comes from Bruce Clay’s early-2000s SEO methodology, which described two variants — a physical silo (URL-folder isolation) and a virtual silo (the same isolation enforced through internal-link structure alone, without moving URLs) — paired with a strict no-cross-linking rule. Later frameworks (HubSpot’s topic clustersSite architecture is how a website's pages are organized, categorized, and interlinked. It controls how crawlers discover pages, how link equity flows, and how clearly search engines understand each page's topical context. Silo structure, hub and spoke, and topic clusters are the three common models., the hub-and-spoke language this cluster uses) cover similar ground — group by topic, link deliberately — but they’re not the same named methodology, and none of them kept the isolation rule. This article uses “silo” the way most practitioners do today: loosely, to mean any topically concentrated hub-and-spoke structure, not Bruce Clay’s original method specifically.

This article assumes you’ve already accepted that a topically concentrated, sensibly cross-linked structure is the goal — and it answers the next question: how do you actually build one? Everything below is a build/audit SOP. Where a “why” question comes up, I’ll point back to the architecture piece rather than re-argue it.

Step 1 — Draw silo boundaries with keyword research

The single most underspecified step in every competing “how to silo” guide is where the boundaries go. Guides say “group by topic” and move on. The useful version is more specific: a subtopic earns its own spoke when it has demonstrated, distinct search demand — not when you need another page to hit a quota.

Practical signals that a subtopic is a real spoke and not a padding page:

• It has its own search volume, separate from the parent term. If “welcome email sequence” gets meaningful searches independent of “email marketing,” it’s a spoke.
• It shows up as its own cluster of related searches / People Also Ask. HubSpot’s own cluster-research mechanic is reusable here even though they don’t call it siloing: start from a seed term and read the People Also Ask boxes — those are the related questions Google already groups with your seed, and each is a candidate spoke.
• Its SERP looks meaningfully different from the parent’s SERP — different intent, different result types, different competing pages. If the two SERPs are nearly identical, you don’t have two topics; you have one.

Ignore the “minimum pages per silo” numbers

You’ll see confident, specific numbers everywhere: 5 pages minimum, 4–8 silos per site, 10–20 cluster pages, HubSpot’s 20 to 30 supporting articles. Notice that they contradict each other and none are sourced to actual data. Treat them as evidence that there is no standard, not as guidance to follow. The real test is distinct demand per subtopic — the same “no magic number of cluster pages” position this cluster’s site-architecture article takes. If a subtopic can’t fill even two or three genuinely different pages, it isn’t a silo; merge it up into the hub.

Watch for cannibalization at the boundary

If two proposed spokes would both reasonably target the same query, the boundary is drawn wrong. That’s not a sign you need two competing pages — it’s a sign those two pages should be one. Drawing boundaries from real keyword clustering (shared SERP overlap, distinct volume, distinct intent) prevents this; drawing them from how your business thinks about its offering doesn’t.

Step 2 — URL structure is organization, not architecture

Here’s the one-line version, because this cluster’s site-architecture article already makes the full case: the link graph is the architecture, not the URL path. Google works out hierarchy from how your pages link to each other, not from your folder names. A page at /email-marketing/welcome-sequences/ signals nothing to Google beyond what its content and its inbound internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. say about it.

So what should you actually do with URLs? Mirror the silo in your folder structure if it helps you and your editors keep the site organized — that’s completely fine and normal. The mistake is believing the folder does the SEO work. And the worst version of that mistake is nofollowing links that cross a folder boundary because the boundary feels sacred (more on that in Step 4). Put pages wherever your CMS is happiest; then go do the actual linking work, which is Step 3.

Step 3 — Build the internal link matrix

This is the load-bearing step, and it’s the one competitors hand-wave. “Link related pages to each other” is not a procedure. Here’s a concrete one, modeled directly on Google’s own description of a topic hierarchy in its ecommerce docs — “add links from menus to category pages, from category pages to sub-category pages, and finally from sub-category pages to all product pages”. That’s Google describing a hub → spoke → sub-spoke link matrixAn internal linking strategy is the deliberate, planned process of deciding which pages on a site should link to which others — with what anchor text and in what priority order — before or while you build, rather than adding links ad hoc as you publish.. Adapt it:

• Hub → every spoke. The hub links to every spoke in its silo — ideally in-context, where that subtopic is introduced in the hub’s body, not just dumped in a list at the bottom. A spoke the hub doesn’t link to is effectively outside the silo.
• Spoke → hub. Every spoke links back to the hub. A “part of our [topic] guide” line, a breadcrumb, or a contextual back-link all work. In the practitioner PageRankPageRank is Google's original recursive link-graph algorithm: a page's score depends on the scores of the pages linking to it, and in the published model each page's score is split across its outbound links (the simplified version: links are weighted votes). Google says it's evolved since launch but still part of its core ranking systems.-flow model of internal linkingLinks between pages on the same site., this is the link that’s supposed to help the hub gather authority from its spokes — Google hasn’t published a formula confirming that, so treat it as a planning heuristic, not a guaranteed ranking effect.
• Spoke → sibling spokes. Spokes link to each other where genuinely relevant to the reader — “for X, see also Y.” Not a forced link to every sibling; not zero either. Relevance is the filter.

The link-matrix spreadsheet

The concrete deliverable nobody in the top results ships: a link-matrix spreadsheet. Rows = every page in the silo. Columns = every page in the silo. Each cell = a checkmark if the row page links to the column page. It looks like this:

Fill it in twice. Once to plan the silo before you write — it makes gaps obvious at a glance (an empty row is a page that isn’t linking out; an empty column is a page nothing points to). Then again after you publish, filled from an actual crawl export instead of from memory, as your audit (Step 6). It’s the same artifact doing double duty, and it’s the single most practical tool in this whole guide.

Step 4 — Reasonable cross-linking rules of thumb

The strict-silo crowd treats every cross-topic link as a leak to be prevented. That instinct is wrong (again, the full argument is in the site-architecture article). Here’s the practical replacement — how to decide when to link out of the silo:

• The “would this link exist anyway?” test. Would you add this link if silos didn’t exist at all — if it were just a helpful pointer for the reader? If yes, add it. If you’re only adding it to “connect the silos,” skip it; if you’re only avoiding it to “protect” the silo, add it.
• Relevance beats direction. Don’t reserve cross-links for “sending equity out,” and don’t avoid them to “keep equity in.” A contextually justified link is good regardless of which silo it points to.
• Volume matters. A handful of well-placed, genuinely relevant cross-links per page is healthy. A page that cross-links constantly to unrelated clusters is diluting its own topical focus — but that’s a content-focus problem, not a reason to reinstate silo walls.

Never nofollow a cross-silo link

This is the build-mechanics myth worth busting directly, because it’s the single most common bad advice in the competing SERP. Bruce Clay’s originating methodology literally recommends it — “If you absolutely had to link the creamy peanut butter page to the flavored jelly page, you would want to do it with a rel='nofollow' link attribute”. Don’t.

PageRankPageRank is Google's original recursive link-graph algorithm: a page's score depends on the scores of the pages linking to it, and in the published model each page's score is split across its outbound links (the simplified version: links are weighted votes). Google says it's evolved since launch but still part of its core ranking systems. sculpting via nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing. has been broken since 2009. Before then, nofollowing a link redistributed its share to the page’s other links; Google changed that so the equity assigned to a nofollowed link now evaporates — it doesn’t get redirected or saved, it’s just gone. So nofollowing a cross-silo link doesn’t protect anything; it destroys the equity that link would have passed. This site’s internal links article covers the sculpting-is-dead history in full. The takeaway for silo-building: link normally when it’s relevant, follow included.

Step 5 — Common implementation mistakes

The four ways silo builds go wrong in practice:

• Over-siloing. Creating more, thinner silos than the topic supports — padding to hit an arbitrary page count. The result is thin, redundant pages competing with each other (cannibalization) instead of pooling authority. Fix: draw boundaries from demand (Step 1), and merge subtopics that can’t stand alone up into the hub.
• Orphaning genuinely cross-topic content. A piece that legitimately spans two silos — say, “email marketing for ecommerce” — gets forced into one silo and either never linked from the other, or awkwardly left out because it “doesn’t fit” the isolation model. Fix: let the page live in one place but be linked from both relevant hubs and spokes. That’s cross-listing, not duplication — this very site does it with an alsoIn pattern (described in the site-architecture article), where an article lists in more than one cluster instead of being walled into one.
• Boundaries that match your org chart, not search intent. Drawing silo lines around how the business organizes its offering (by product line, by internal department) instead of how searchers group and phrase queries. Fix: validate every boundary against actual keyword/SERP clustering (Step 1), not an internal structure.
• Trusting the folder instead of the link graph. You put the pages in the /topic/ folder, assume the silo is “built,” and never check whether the pages actually link to each other. The folder existing doesn’t mean the linking work got done — which is exactly why Step 6 exists.

Step 6 — Audit that the link concentration actually happened

You planned the matrix in Step 3. Now verify it landed. Three complementary tools: none of them, alone, proves the topics “worked” for a reader or a search engine — a visualization or a matrix fill rate is a proxy, not a score. Beyond hub↔spoke links, check whether every spoke is actually reachable by following links from the hub (not just present in a report), whether the cross-group edges to other silos are the ones you intended and not stray leftovers, and then walk the pages yourself the way a reader would — a crawl finds link edges, it doesn’t tell you whether the path makes sense to a person. Track discovery, crawl, and indexStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. status for the silo’s pages separately from any ranking change that follows; a rewired link graph can move all three, but that’s not evidence the rewiring caused a ranking or traffic change — too many other things move at once on a live site to claim that.

Screaming Frog crawl visualisations

Crawl the site and open the force-directed diagram or tree graph — a node-and-line renderingTurning HTML, CSS, and JavaScript into the final visual page and DOM. of the actual crawled link graph, colored by crawl depthCrawl depth usually means click depth — how many clicks it takes to reach a page from the homepage by following internal links. It can also mean a crawler setting that limits how many levels deep a crawl goes before it stops.. It’s the fastest way to see whether a topic’s pages visually cluster together or are scattered and disconnected. Use it as a communication and pattern-spotting tool, not a data source — Screaming Frog is explicit that “the crawl visualisations are useful when analysing site architecture, and internal linking”, but also warns they “don’t provide any more data than is already available in a crawl… don’t always tell the whole story.” Great for spotting a stranded spoke at a glance; not a substitute for the data-level checks below.

Ahrefs Site Audit

Three reports, in the same priority order this cluster’s internal-links article uses for any internal-link audit (orphans → broken links → link-equity opportunities):

1. Orphan PagesAn orphan page is a page on your site that no other page links to internally. Because crawlers discover pages by following links, an orphan page is effectively invisible to search engines unless it's in an XML sitemap or linked from an external site. — zero-inlink pages. A silo build most commonly leaves behind the last spoke you added, so this catches the pages nothing in the silo points to.
2. Internal Link Issues — broken internal links inside the silo. A hub linking to a spoke’s old URL is a link that isn’t doing its job.
3. Link Opportunities — “relevant internal linking suggestions” based on keyword overlap between pages, sortable by Page Rating so you can prioritize which high-authority pages should be doing the linking. I’ve written about using this for silo linking specifically in SEO Silo Structure: Why It Makes No Sense (And What to Do Instead) — “the Link Opportunities tool… suggests where you should add internal links.”

The manual link-matrix audit

The spreadsheet from Step 3, filled from reality. Crawl the silo’s URLs, export the outlinks (Screaming Frog’s outlinks export, or an Ahrefs crawled-pages export), and fill each cell from actual crawl data — not from what you think you linked. Empty rows = pages not doing their linking job. Empty columns = pages nothing in the silo points to (an orphan risk within the silo even if the page is linked from somewhere else on the site).

TIP Prioritize the missing links after the matrix audit

The score and visible reasons make a proposed link reviewable. They do not prove that the source passage supports the destination, so keep editorial relevance as the final gate.

Score supplied hub and spoke pages with my free Internal Link Cluster Visualizer Free

1. Supply the silo pages with titles, URLs, and their current internal links.
2. Review high-scoring pairs where the target has few or no supplied links.
3. Add only links that fit the source passage, then refill the matrix from a new crawl.

Tap to enlarge

The row scores a proposed link at 76.6 from Internal linking explained to Topic cluster strategy. The visible reasons include shared terms, a matching blog URL segment, and a target with no supplied links or few supplied inlinks.

When to re-run it

After the initial build, then every time you add a spoke to an existing silo. New pages are the most common point of failure — you’ll update the hub to link to the new spoke and forget the sibling cross-links, or forget to link the new spoke back to the hub. Re-running the matrix after each addition catches that in about five minutes.

FAQs

How many pages should a silo have? No fixed number. Competing guides cite 5, 8, 10–20, and 20–30 and contradict each other. Use distinct search demand per subtopic as the test, not a quota.

Should I nofollow links between silos? No. PageRank sculpting via nofollow has been broken since 2009 — the equity evaporates rather than being protected. Link normally when it’s relevant.

Does my URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly. need to match my silo structureA silo structure is a site organized by grouping related content into topic areas and concentrating internal links within each topic so a hub page and its supporting pages reinforce each other. Modern practice keeps the topical concentration and drops the old rule against cross-linking between topics.? It can, for organizational convenience, but Google reads your internal link graph, not your folder names. See this cluster’s site-architecture article for the full case.

What tools audit a silo structure? Screaming Frog crawl visualisations, Ahrefs Site Audit (Orphan Pages, Internal Link Issues, Link Opportunities), or a manual link-matrix spreadsheet built from a crawl export.

Can a page belong to more than one silo? Yes — via cross-listing (linking it from multiple relevant hubs) rather than duplicating it or forcing it into one bucket. This site’s alsoIn pattern is a working example.

AI summary

A condensed take on the Advanced version:

• This is a build SOP, not a debate. Whether strict silos are good is settled in this cluster’s site-architecture article; here the goal (topical concentration + sensible cross-linking) is assumed and the question is how to build it. (“Silo” itself traces to Bruce Clay’s physical/virtual isolation methodology — this article uses the term loosely, without the isolation rule.)
• Step 1 — boundaries from demand. A subtopic earns a spoke when it has distinct search volume, its own related-search/PAA cluster, and a meaningfully different SERP from the parent. Ignore the contradictory “5/8/10–20/20–30 pages” numbers — they’re evidence there’s no standard. Watch for cannibalization at the boundary.
• Step 2 — URLs are organization, not architecture. Mirror the silo in folders for CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. sanity if you want, but the link graph is the architecture; the folder doesn’t do the SEO work.
• Step 3 — wire the matrix. Hub → every spoke (in-context), spoke → hub, spoke → relevant siblings. Use a link-matrix spreadsheet (rows/columns = pages, cell = linked?) to plan it and later audit it.
• Step 4 — cross-link sensibly. Add the link if it’d exist anyway; relevance beats direction; a few per page is healthy. Never nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing. a cross-silo link — sculpting died in 2009 and the equity evaporates.
• Step 5 — avoid the classic mistakes: over-siloing/padding, orphaning cross-topic content (fix with cross-listing, not duplication), boundaries drawn on the org chart instead of search intent, and trusting the folder over the link graph.
• Step 6 — audit with a crawl: Screaming Frog visualisations to see clustering, Ahrefs Site Audit (orphans → broken links → link opportunities), and the manual link-matrix filled from crawl data. Check reachability and cross-group edges too, walk the pages like a reader, and track discovery/crawl/indexStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. separately from ranking — none of it proves the rewiring caused a ranking change. Re-run every time you add a spoke.

Official documentation

Primary-source guidance most relevant to the build mechanics — wiring a topic hierarchy and keeping every page linked. (The broader silo-vs-cluster model debate is documented in this cluster’s site-architecture and website-structure articles; those cite Google’s pyramid/hub-page language and Bing’s click-depth framing, so I won’t repeat them here.)

Google

• Help Google understand your ecommerce site structure — the closest thing Google has to a “how to wire a topic hierarchy” doc; describes the menu → category → sub-category → product link matrixAn internal linking strategy is the deliberate, planned process of deciding which pages on a site should link to which others — with what anchor text and in what priority order — before or while you build, rather than adding links ad hoc as you publish. directly.
• Crawlable links — Make your links crawlable — every page you care about should have a link from at least one other page (the anchor for “don’t orphan cross-topic content”), plus Google’s anchor-text guidance for writing hub→spoke and spoke→spoke link textAnchor text is the visible, clickable text of a hyperlink. It tells readers what they'll find on the other end and gives search engines context about the linked page..

Bing / Microsoft

• Bing Webmaster Guidelines — a clear, crawlable, hierarchical category/subcategory structure with reasonable internal linkingAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them.. (Bing is more prescriptive on click depthCrawl depth usually means click depth — how many clicks it takes to reach a page from the homepage by following internal links. It can also mean a crawler setting that limits how many levels deep a crawl goes before it stops. than Google; this cluster’s website-structure article covers Bing’s three-click framing.)

Quotes from the source

On-the-record statements relevant to building — and to the myth about walling silos off. (The Google/Bing rep quotes about whether silos are good — Mueller’s pyramid line, the folder-structure line — live in this cluster’s site-architecture article, already deep-linked there; I won’t duplicate them.)

Google Search Central docs — wiring the hierarchy (ecommerce site structureWebsite structure (site architecture) is a site's visible hierarchy, navigation, breadcrumbs, and URL organization — how pages relate and how people and search engines move between them. Internal linking is the primary signal Google reads to understand that structure, not URL folders.)

• “add links from menus to category pages, from category pages to sub-category pages, and finally from sub-category pages to all product pages.” Jump to quote

Google Search Central docs — don’t orphan a page (crawlable links)

• “Every page you care about should have a link from at least one other page on your site.” Jump to quote

Bruce Clay — the outdated nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing.-the-cross-silo-link advice (cited critically)

• “If you absolutely had to link the creamy peanut butter page to the flavored jelly page, you would want to do it with a rel='nofollow' link attribute.” Read the source — this is the myth this article busts: sculpting via nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing. has been dead since 2009, so the equity evaporates rather than being protected.

Screaming Frog — what crawl visualisations are (and aren’t) for

• “the crawl visualisations are useful when analysing site architectureSite architecture is how a website's pages are organized, categorized, and interlinked. It controls how crawlers discover pages, how link equity flows, and how clearly search engines understand each page's topical context. Silo structure, hub and spoke, and topic clusters are the three common models., and internal linkingAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them..” Read the source
• “They don’t provide any more data than is already available in a crawl. They are often not the best way to diagnose issues and don’t always tell the whole story.” Read the source

Standard operating procedure: build a topic silo

Run this end to end for a new silo, or from Step 4 to retrofit an existing group of pages.

1. Pick the hub topic. Choose a broad topic with real breadth and search demand that will anchor the silo. This becomes your hub/pillar page.
2. Cluster the keywords. Pull the seed term’s related searches, People Also Ask, and keyword-tool suggestions. Group them by intent. Each group with distinct search demand and a distinct SERP is a candidate spoke.
3. Cut the padding. For each candidate spoke, confirm it has demand separate from the parent. If it can’t fill 2–3 genuinely different pages, merge it up into the hub instead of creating a thin page. Kill any two candidates that would cannibalize the same query — make them one.
4. Lay out URLs (optional, cosmetic). Mirror the silo in a /topic/ folder if it helps your CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms./editorial organization. Do not treat this as the SEO work.
5. Draft the link-matrix plan. Build the rows-and-columns spreadsheet (every page in the silo on both axes) and mark the links you intend: hub → every spoke, every spoke → hub, spoke → each genuinely relevant sibling.
6. Write and wire. Produce the hub and spokesSite architecture is how a website's pages are organized, categorized, and interlinked. It controls how crawlers discover pages, how link equity flows, and how clearly search engines understand each page's topical context. Silo structure, hub and spoke, and topic clusters are the three common models.. Place hub→spoke links in context where each subtopic is introduced, not in a bottom-of-page dump. Add the spoke→hub and spoke→sibling links per your matrix.
7. Add sensible cross-links. For each page, add out-of-silo links that pass the “would this link exist anyway?” test. Follow, never nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing..
8. Crawl and verify. Crawl the silo (Screaming Frog / Ahrefs Site Audit). Fill the matrix from the crawl’s actual outlinks. Fix empty rows (pages not linking out) and empty columns (pages nothing points to). Clear orphan-pages and broken-internal-link reports.
9. Set a re-audit trigger. Every time you add a spoke later, re-run Steps 5–8 for that page — update the hub, add the sibling cross-links, re-crawl.

Silo build & audit checklist

Work it top to bottom; the boundary work up front prevents the linking rework later.

Boundaries (Step 1)

• Each spoke has distinct search demand separate from the hub term.
• Each spoke has its own related-search / People Also Ask cluster.
• Each spoke’s SERP looks meaningfully different from the parent’s.
• No spoke exists only to hit a “minimum pages” number.
• No two spokes target the same query (no boundary cannibalization).

Structure (Steps 2–3)

• Folder/URL layout (if used) is treated as CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. convenience, not as the silo.
• The hub links to every spoke, in-context, not just a footer list.
• Every spoke links back to the hub.
• Spokes link to genuinely relevant sibling spokes.
• A link-matrix spreadsheet exists (rows/columns = pages, cells = links).

Cross-linking (Step 4)

• Cross-silo links exist where a reader would genuinely follow them.
• Each cross-link passes the “would this exist if silos didn’t?” test.
• No internal linkAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. — cross-silo or otherwise — is nofollowed to “protect” equity.

Mistakes to rule out (Step 5)

• Not over-siloed (no thin, redundant, cannibalizing spokes).
• Cross-topic content is cross-listed (linked from both hubs), not orphaned.
• Boundaries match search intent, not the org chart.
• You verified the actual link graph, not just the folder.

Audit (Step 6)

• Screaming Frog visualisation shows the silo’s pages clustering, not scattered.
• Ahrefs Site Audit: no orphan pagesAn orphan page is a page on your site that no other page links to internally. Because crawlers discover pages by following links, an orphan page is effectively invisible to search engines unless it's in an XML sitemap or linked from an external site. in the silo.
• Ahrefs Site Audit: no broken internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. inside the silo.
• Matrix filled from a real crawl export — no empty rows or unexpected empty columns.
• Re-audit scheduled for the next time a spoke is added.

The link-matrix template

This is the artifact the competing guides don’t ship. Rows are the “linking from” page; columns are the “linking to” page; each cell is ✓ if that link exists. Build it in any spreadsheet — one tab to plan, one filled from a crawl to audit.

How to read it

• Hub row should be all ✓ — the hub links to every spoke. A gap here means a spoke that isn’t really in the silo.
• Hub column should be all ✓ — every spoke links back to the hub. A gap here is a spoke that, in the practitioner link-flow model, isn’t returning authority to the hub — a planning signal to fix, not a measurable Google score.
• A spoke’s row being empty (besides the hub) = a page that isn’t linking to any siblings; look for a relevant sibling to add.
• A spoke’s column being sparse = few pages point to it; it’s under-linked and at orphan risk within the silo.

To turn it into an audit: crawl the silo’s URLs, export outlinks (Screaming Frog outlinks export or an Ahrefs crawled-pages export), and fill the cells from the real data instead of from memory. The gaps between your plan tab and your crawl tab are your to-do list.

Cross-links live outside this matrix by design — the matrix is for within-silo completeness. Track out-of-silo links separately (a simple “page → external-to-silo target → reason” list) so you can sanity-check they each pass the “would this exist anyway?” test.

The mental models

1. Boundaries follow demand, not a quota. A subtopic earns a spoke when it has distinct search demand (its own volume, its own PAA cluster, its own SERP). “5 pages minimum” and its cousins are folklore that contradict each other — the demand test is the only one that holds.

2. The folder is a filing cabinet; the links are the architecture. Mirror the silo in folders for your own sanity, but the internal linkAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. graph is what search engines read. Build and audit the link graph, not the folder tree.

3. The matrix is the silo. A silo is complete when the hub links to every spoke, every spoke links back, and spokes link to relevant siblings. Rows and columns on a spreadsheet make “complete” checkable — plan it there, verify it there.

4. Cross-links are features, not leaks. The test is “would I add this link if silos didn’t exist?” If yes, it belongs. The instinct to wall silos off — especially by nofollowing cross-links — is the exact outdated practice that destroys equity rather than protecting it.

5. Trust the crawl, not your memory. You think you linked everything; the crawl knows whether you did. The build isn’t done when the pages are published — it’s done when a crawl-filled matrix has no unexpected gaps.

Should this subtopic get its own spoke — or a cross-link — or nothing?

The two hardest calls in a silo build are where the boundaries go and when to link across them. Walk a candidate subtopic (or a candidate cross-link) through this.

Silo-building cheat sheet

The link matrixAn internal linking strategy is the deliberate, planned process of deciding which pages on a site should link to which others — with what anchor text and in what priority order — before or while you build, rather than adding links ad hoc as you publish. (within a silo)

Boundary test — is it a real spoke?

• Distinct search volume vs. parent? ✓ needed
• Own related-search / PAA cluster? ✓ needed
• SERP differs from parent’s? ✓ needed
• Only exists to hit a page count? → merge up, not a spoke
• Same query as another spoke? → combine, don’t compete

The “pages per silo” numbers, decoded

• 5 minimum / 4–8 silos / 10–20 pages / 20–30 articles → all folklore, all contradict each other. Use demand, not a quota.

Myths to drop

• “The folder builds the silo.” → No; the link graph is the architecture.
• “Never link across silos.” → Link when relevant.
• “Nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing. cross-silo links to protect equity.” → Dead since 2009; equity evaporates.
• “A silo is a ranking factor.” → No silo detector; it’s a method for organizing ordinary internal-link signals.

Audit tools

• Screaming Frog visualisations → see whether pages cluster (pattern-spotting, not data).
• Ahrefs Site Audit → Orphan PagesAn orphan page is a page on your site that no other page links to internally. Because crawlers discover pages by following links, an orphan page is effectively invisible to search engines unless it's in an XML sitemap or linked from an external site. → Internal LinkAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. Issues → Link Opportunities.
• Link-matrix spreadsheet → filled from a crawl export → find empty rows/columns.
• Re-audit → every time you add a spoke.

Silo anti-patterns — and the fix

Over-siloing to hit a page count. Symptom: Thin, near-duplicateThe same or very similar primary content reachable at more than one URL. There's no general duplicate content penalty — the real costs are possible signal dilution, the wrong URL getting chosen, and less-efficient crawling. spokes invented to reach “5 pages minimum,” competing with each other for the same query. Fix: Draw boundaries from distinct demand; merge subtopics that can’t stand alone up into the hub.

Nofollowing cross-silo links. Symptom: rel="nofollow" on internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. that cross a folder boundary, added to “keep equity in the silo.” Fix: Remove it. Sculpting has been broken since 2009 — the equity evaporates instead of being redirected. Link followed when it’s relevant.

Building the folder and calling it done. Symptom: Pages sit in /topic/ but nothing inside the folder links to anything else inside it; the “silo” exists only as a URL path. Fix: Do the link-matrix work (Step 3) and verify it with a crawl (Step 6); the folder is filing, not architecture.

Orphaning cross-topic content. Symptom: A piece that spans two silos gets forced into one and never linked from the other — or dropped entirely because it “doesn’t fit.” Fix: Cross-list it — link it from both relevant hubs/spokes (the alsoIn pattern), one page, two homes.

Org-chart boundaries. Symptom: Silos drawn around product lines or internal departments, so the structure matches how the company thinks, not how searchers query. Fix: Re-cluster the boundaries from real keyword/SERP data.

Bottom-of-page link dumps. Symptom: The hub “links to every spoke” only via a related-links list at the very bottom, with no in-context links. Fix: Place hub→spoke links where each subtopic is introduced in the hub’s body — contextual links carry more signal and get followed by readers.

Set-and-forget after launch. Symptom: A new spoke gets added months later; the hub never gets the new link and the sibling cross-links never happen. Fix: Re-run the link matrixAn internal linking strategy is the deliberate, planned process of deciding which pages on a site should link to which others — with what anchor text and in what priority order — before or while you build, rather than adding links ad hoc as you publish. for every new spoke — it’s the single most common point of silo decay.

Review proposed silo boundaries

Paste a list of pages with target topic, intent, and proposed hub into this prompt:

Review this proposed topic-silo map as an information architecture, not as a URL
folder exercise. Flag overlapping hubs, spokes that belong under a different hub,
missing subtopics supported by the supplied data, and cross-topic links that would
help a reader. Do not invent keywords or search volume. Return: boundary decisions,
hub-to-spoke links, spoke-to-hub links, useful cross-links, and unresolved cases.

Turn a crawl export into a link-gap review

Paste source URL, destination URL, anchor textAnchor text is the visible, clickable text of a hyperlink. It tells readers what they'll find on the other end and gives search engines context about the linked page., page topic, and crawl-depth columns:

Compare the observed internal-link graph with the supplied hub-and-spoke plan. Find
missing hub-to-spoke and spoke-to-hub links, weak or ambiguous anchor text, isolated
spokes, and relevant cross-silo links that the strict plan omitted. Separate facts
visible in the data from recommendations. Do not recommend nofollow to preserve a
silo boundary.

Verify the planned link matrix shipped

Test to run: Crawl the site and compare observed source-destination pairs with the approved hub-and-spoke matrix. Expected result: Every published spoke has a crawlable link from its hub and links back to the correct hub. Failure interpretation: A template, deployment, or mapping omission left the planned edge out of the graph. Monitoring window: Immediate after the deployment crawl. Rollback trigger: The release removes existing useful links or routes a cluster to the wrong hub.

Verify cross-links remain intentional

Test to run: Export links crossing topic groups and review their source context and anchor textAnchor text is the visible, clickable text of a hyperlink. It tells readers what they'll find on the other end and gives search engines context about the linked page.. Expected result: Each cross-link serves a related reader task; there are no blanket nofollow rules or accidental sitewide cross-links. Failure interpretation: The implementation either enforced artificial isolation or added links without topical context. Monitoring window: After each template or navigation release. Rollback trigger: A global rule suppresses legitimate links or introduces a large unrelated sitewide link set.

Test yourself: building a silo structure

Five quick questions on the silo build-and-audit process. Pick an answer for each, then check.

Resources worth your time

My related writing

• SEO Silo Structure: Why It Makes No Sense (And What to Do Instead) — the Ahrefs take on strict silos, including using the Link Opportunities tool to find where to add internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them..
• Internal Links for SEO: An Actionable Guide — how internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. pass PageRankPageRank is Google's original recursive link-graph algorithm: a page's score depends on the scores of the pages linking to it, and in the published model each page's score is split across its outbound links (the simplified version: links are weighted votes). Google says it's evolved since launch but still part of its core ranking systems. and shape the link graph that is the silo; the audit workflow this article’s Step 6 builds on.
• The Beginner’s Guide to Technical SEO — where site structureWebsite structure (site architecture) is a site's visible hierarchy, navigation, breadcrumbs, and URL organization — how pages relate and how people and search engines move between them. Internal linking is the primary signal Google reads to understand that structure, not URL folders. and internal linkingLinks between pages on the same site. fit in the bigger picture.

My speaking

• Local Link Building (Pubcon Local 2021) — my internal-linking talk; the closest adjacent material on wiring links deliberately (internal links seemed to matter proportionally more for local queries).

From around the industry

• SEO Silos: How to Build a Website SEO Silo (Bruce Clay) — the originating strict-silo methodology, cited here critically for its outdated nofollowrel=\"nofollow\" is a value of the HTML link rel attribute that tells search engines you don't vouch for a linked page and don't want to pass ranking signals to it. Since 2019–2020 Google treats it as a hint, not a directive — and it does not reliably block crawling or indexing.-the-cross-silo-link advice.
• Screaming Frog — Site Architecture & Crawl Visualisations — how to see whether a topic’s pages cluster, and the caveat that visualisations don’t tell the whole story.
• Ahrefs Academy — Link Opportunities — the Site Audit report that surfaces relevant internal-linking suggestions, sortable by Page Rating.
• Topic Clusters: The Next Evolution of SEO (HubSpot) — the People Also Ask cluster-research mechanic reused in Step 1 (and the source of the “20 to 30 supporting articles” number cited as folklore).
• What Is Silo SEO and Is There a Better Option? (Semrush) — a mainstream silo walkthrough for comparison.
• SEO Silo Structure (Ahrefs glossary) — a short reference definition of the silo model and the objection to its no-cross-linking rule.

Silo Structure

A silo structure is a site organized by grouping related content into topic areas and concentrating internal links within each topic so a hub page and its supporting pages reinforce each other. Modern practice keeps the topical concentration and drops the old rule against cross-linking between topics.

Related: Internal Link, Site Architecture, Anchor Text

Silo Structure

A silo structure (also called website silo architecture or just siloing) is a way of organizing a site’s content by topic and concentrating internal linksAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. within each topic group, so that a hub page and its supporting pages reinforce one another rather than being scattered across the site with no deliberate link pattern. Siloing is really a build process — the practical set of steps you follow to produce that topically concentrated internal linkingLinks between pages on the same site..

The term comes from Bruce Clay’s early-2000s “SEO silo” methodology, which paired topical grouping with URL-folder isolation and a strict rule against linking between silos. Modern practice — reflected in Google’s own guidance, HubSpot’s “topic cluster” model, and Ahrefs’ own position — keeps the topical-concentration idea and drops the isolation rule. A silo done well today is functionally a hub-and-spoke or topic-cluster structure: a pillar/hub page linking to every spoke, spokes linking back and to relevant siblings, and deliberate, limited links out to genuinely related content elsewhere on the site.

Two things a silo structure is not. It’s not a URL-folder trick — search engines read your internal linkAn internal link is a hyperlink from one page on a website to another page on the same website. Internal links help search engines discover your pages and pass ranking signals (PageRank and anchor-text context) between them. graph, not your path segments, so putting pages in a folder doesn’t build the silo. And it’s not a ranking factor Google checks for — there’s no “silo detector.” What actually gets evaluated is ordinary internal-linking signals (discovery, context, and inferred importance); siloing is simply a method for organizing those signals well.

Related: Internal Link, Site Architecture, Anchor Text

Changelog