Table of Contents
TL;DR
- A small website’s link problem is rarely about how many links it has. The real issue is how few of those links connect related posts.
- Aim for two to three internal links in the body of every post, pointing to topically-related posts on the same site. The practitioner band is two to five links per thousand words of body content.
- Pages with zero incoming internal links are orphan pages. Crawlers struggle to find them and AI chat surfaces rarely cite them. Fixing the orphans is one of the highest-impact interventions on most small sites.
- Anchor text matters as much as link count. A descriptive three-to-six-word phrase that names the linked topic outperforms generic placeholders by a wide margin.
- Open the post earning the most impressions in Search Console. Add two to three links from it to topically-related older posts. Repeat next month. One year of that habit covers most of what a small site needs.
Fewer than six percent of internal links on a typical website reach pages buried four clicks from the homepage.
That single number explains why a small website with five thoughtful blog posts often performs worse than a competitor with three. Five posts and no links between them are five pages standing alone. Three posts with links between them are a small cluster. Both Google and the AI chat surfaces read clusters more cleanly than they read isolated pages.
The good news for a small website is that internal linking is the one part of SEO where the work scales down well. A site with twelve posts can be fully linked in one afternoon. A site with three hundred posts cannot, because the candidate pairings grow faster than the time to review them. The small site is at a structural advantage here.
This piece walks through the simple internal-linking rules that earn the credit on a small website in 2026. The two-to-three-link target per post. The orphan-page check that takes ten minutes.
The anchor-text habit that quietly compounds. And the single one-hour intervention worth doing first.
Why does internal linking matter more for a small website in 2026?
Internal linking gives both Google and the chat surfaces a way to see how the posts on your site relate to one another.
A small site with five disconnected posts looks like five separate pages. The same five posts with two or three links between them looks like a cluster on a topic. Both engines read clusters more cleanly than they read isolated pages. The cluster shape signals that the site treats the topic seriously enough to cover it more than once.
The 2026 reading is straightforward. Google still uses internal links to distribute crawl attention and to pass relevance signals between pages. The AI chats use the link graph as one input when deciding which page on a topic to cite.
A page nobody links to from elsewhere on the same site signals something to both engines. Even the site itself does not consider the page core to its topic.
Five posts and three internal links between them outperform fifty posts with no internal links between them, all else equal. The cluster shape is what the engines reward. The count of posts is secondary.
The site article on building topical authority on a small website covers the cluster idea from the topic side. This piece covers the same cluster from the linking side.
What is the simplest internal-linking rule a small website can follow?
Every new post gets two to three links out to older posts on related topics.
Every older post gets one new link added back when the new post publishes. The practitioner band is two to five contextual internal links per thousand words of body content. Below the band a post sits orphaned. Above it the page begins to read like a link directory rather than an article with a purpose.
The two-to-five band comes from practitioner consensus, not from a Google-published number. Google has acknowledged that very high internal-link counts on a single page can dilute the relevance signal each link passes. A thousand-word post with twenty internal links carries less weight per link than the same post with three.
The practical floor is two. The practical ceiling on a thousand-word post is around five.
A small site does not need a database to track this. A simple habit does the job. When you publish a new post, list three older posts the new one connects to topically.
Add a link from each old post back to the new one. Add three links from the new post out to the topically-related older ones. The whole habit takes ten minutes per publish and quietly builds the cluster shape over time.
How do you find the right pages to link to from a new post?
List three to five topics the new post touches.
Search your own site for older posts on those topics. The fastest free way is the search box on your own site or a free site-search plugin that orders by relevance rather than recency. Read the first three results.
Pick the one most directly related. The candidate is now a link target.
Place the link inside a supporting paragraph after the answer block of the section it best fits. Use a descriptive anchor text that names the linked topic clearly enough that a reader who never hovers the link still understands where it leads. A reader who reads "the answer-block pattern that wins citations" knows where the link goes. A reader who reads "click here for more" does not.
A common trap on a small site is to link only to whichever post happens to be most recent. The recency bias produces a chain rather than a cluster. The reader who follows a 2026 post on internal linking back to a 2026 post on Search Console covers no new ground. Both posts already shared half of the same audience.
Linking to an older post on a topic you have not covered for eighteen months is often more useful. Last week’s adjacent post is the easier link to make and the weaker one for the reader.
Why does descriptive anchor text outrank "click here" and "read more"?
Anchor text is the short clickable phrase the reader sees.
It tells the engine what the linked page is about. A descriptive three-to-six-word phrase that names the topic outperforms generic anchors such as "click here" or "read more" by a wide margin. The generic anchors pass minimal topical signal. The descriptive ones carry the relevance the engines reward and the orientation the reader needs.
The widely cited heuristic among SEO practitioners has four anchor-type bands. Exact-match anchors sit around fifteen to twenty-five percent of the total. Partial-match around thirty to forty percent. Semantic variants around twenty-five to thirty-five.
Generic anchors well under ten. Those ratios are practitioner consensus, not Google-published thresholds. The honest framing is that variation prevents the over-optimization signal a hundred identical anchors send, and that descriptive beats generic at every ratio.
The simplest test is whether the anchor text alone tells the reader what the linked page covers. Strip the surrounding sentence in your head and read the link text by itself. If it still names what the link points to, the anchor is doing its job. If it reads as a placeholder, rewrite the anchor before the post ships.
Where should the link sit within a section so it does not break the citation?
Place internal links inside the supporting paragraphs that follow the answer block, never inside the answer block itself.
The 40-to-60 word answer block is what the engines lift cleanly when they cite a page. A link inside that block signals that the real answer lives elsewhere, which is the opposite of what the answer block is supposed to demonstrate. The lifted passage is also harder to render cleanly on the citing surface when it carries an inline link. Both reasons push the link outside the block.
The Indig analysis of 1.2 million AI chat responses (February 2026) showed a clear pattern. Pages whose answer blocks carry no inline links cite at meaningfully higher rates than pages whose answer blocks carry one or more. The reason is mechanical.
The chat surface lifts the block as a passage. The link inside that passage breaks the passage’s claim to be the answer.
The supporting paragraph after the answer block is the natural home for the link. The reader who has read the answer block already has the answer. The link in the supporting paragraph sends the curious reader to deeper context.
The site article on structuring content for featured snippets in 2026 covers the answer-block shape in detail. This piece treats the answer block as the constraint the link placement has to respect.
What are orphan pages and how do you find yours?
An orphan page is one that no other page on your site links to.
It receives no relevance signal from the rest of the site. Both crawlers and AI chat surfaces struggle to discover it beyond the sitemap. Most small sites have at least one or two orphan pages.
A site of any age that has never run an orphan check usually has a handful. The orphans are often older posts that lost their menu placement during a redesign.
The free tier of Screaming Frog covers a site of up to five hundred URLs and surfaces orphan pages directly through the Reports menu. The check works once you connect either Google Search Console or an uploaded XML sitemap. The free Ahrefs Webmaster Tools also flags orphan pages within the Site Audit issues tab. Either tool finds the orphans on a small site in under ten minutes.
Once the orphans are listed, each one needs a triage call. A page with no traffic and content superseded by a newer post is best redirected to the replacement. A page with content that still belongs on the site needs at least two internal links added from older posts that share its topic. A page with thin or duplicate content is often a delete-or-merge case rather than a re-link case.
A widely cited Botify case showed a production site where seventy percent of pages crawled by Google were orphaned relative to the internal link structure. The number is striking but should not be read as an industry benchmark. It is one site, not the average.
The point is that orphan pages quietly accumulate on any site live for more than a year. A small site benefits from a quarterly orphan check more than from a stack of fancier audits.
How many internal links is too many on one page?
The practitioner heuristic is two to five contextual internal links per thousand words of body content.
Total per-page link count, including navigation, sidebar, and footer, stays under one hundred and fifty. Above the upper band the relevance signal each link passes thins out. The page begins to read like a link directory rather than an article with a purpose. Below the lower band the post sits underconnected to the rest of the site.
Google has acknowledged through John Mueller that pages carrying very high numbers of internal links can dilute the relevance signal each link passes. The cited threshold among practitioners is around one hundred or more. The pattern risks reading as manipulative when it repeats site-wide.
The original quote is reported by practitioners — no Google-published number replaces the heuristic. Treat one hundred and fifty as the practical ceiling and stop counting once the body sits inside the two-to-five-per-thousand band.
The sidebar and the footer are where the count silently grows. A "recent posts" widget that shows ten posts in the sidebar adds ten internal links to every page on the site. The same widget shown on every page of a hundred-page site means a thousand identical links across the corpus.
Trim the sidebar to three or four links that actually help the reader. The body links then carry more of the relevance signal that matters for both engines and AI chat citations.
Other questions worth answering
What happens to the cluster shape when you delete an older essay?
A 404 dead end is what every outbound URL becomes when you delete the target older essay. The clean fix is to redirect the deleted URL to its closest surviving relative. Without that step, the cluster develops orphaned threads across your archive. The 2026 reality is that even a few stale 404s on a modest property meaningfully erode trust signals.
How do categories and tags contribute to your cluster shape?
WordPress generates a category hub for each label you assign to your essays. A reader who clicks a category sees every URL filed under it, giving the engines a second route into older essays. Tags work similarly but at a finer grain. As of 2026, Google and Perplexity both treat well-named categories as topical containers worth surfacing.
When does a redirect beat a re-connection for an older URL?
Two signals push you toward a redirect over a re-connection. First, the older URL has no impressions over the past quarter and a newer 2026 essay already covers the same ground. Second, the topic context is genuinely obsolete, as with a years-old walkthrough of a Google feature that no longer exists. Otherwise the older URL still belongs on the domain.
What does crawl depth mean for URLs buried far from the homepage?
Crawl depth counts the clicks a crawler takes from the homepage to a URL. An industry measurement of 2.5 million in-body connections across 1,700 sites in 2025 found that under six percent reach URLs at depth four or deeper. Pulling a buried URL up to depth two via a strong hub reference is among the highest-impact moves on a modest property.
Where should you add the first three internal links if you have one hour?
Open Google Search Console.
Sort pages by impressions for the last twenty-eight days. Open the page with the highest impressions. Add two to three internal links from it to topically-related older posts on the site.
Repeat on the next two highest-impression pages. One hour, three pages, six to nine new links, and the cluster on your strongest topic begins to read as a cluster.
The pages earning the most impressions are the ones the engines have already decided are worth showing. Each of those pages is a hub in the small-site link graph, even if the site owner never thought of it that way. A link from a high-impression page to a related page passes more relevance signal than a link from a low-impression page does. The high-impression page is also the one already attracting the readers who would benefit from the related content.
The repeat-monthly habit is what compounds. A small site with twelve posts and one hour of internal linking per month becomes a fully linked twelve-post cluster in three months. Add one new post each month.
Add one new round of links from the existing posts to the new one. The cluster grows without ever needing a separate "internal linking project" on the calendar.
If you have one hour and want a second opinion on which pages to link first, you can contact me here. Send the URL of the site and the top three impression pages from Search Console. I will name the two or three links I would add first. No pitch.