Internal links and orphan pages: passing backlink value to the pages that need it

An internal link is a link from one page of a site to another page of the same site. An orphan page is a page that has no internal links pointing to it. PageRank sculpting is an old tactic of adding nofollow to some internal links in the hope of sending more value through the rest. All three matter to link building for one reason: a backlink benefits the page it points to, and it is internal links that decide how much of that benefit reaches anything else. Most sites earn their links on a few pages and need rankings on others, so the work of building links is only half finished when the link goes live.

Why backlinks land in the wrong place

The pages that attract links and the pages that make money are usually different pages. People link to homepages when they mention a brand, and to guides, research, tools and news when they need a source. They rarely link to a product category, a service page or a pricing page, because there is seldom an editorial reason to. The lesson on which pages need links goes through this pattern by page type. The result is a site whose earned value sits on its homepage and a few informational pages, while the pages that would turn a ranking into revenue have few or no backlinks of their own.

Buying links straight to commercial pages is one answer, and it carries the risks that come with it, since a run of keyword-anchored links to a sales page is the pattern Google’s spam systems are built to notice. The other answer costs nothing and breaks no rule. In the PageRank model a page shares its score among the pages it links to, which is described in the lesson on link equity, and that sharing does not stop at the boundary of your own site. A guide that has earned links from many sites passes part of what it has to every page it links to, including yours. Whether the pricing page gets any of it depends on whether the guide links there, and on how many other links compete for the same share.

What Google says internal links do

Google’s link best practices page gives internal links two jobs. It says Google uses links as a signal when determining the relevancy of pages and to find new pages to crawl, and it states that every page you care about should have a link from at least one other page on your site. The same page asks for anchor text that is descriptive, reasonably concise and relevant to the page it points to, and it applies that advice to internal links as much as to external ones. An internal link reading “click here” tells Google as little about the target as the same anchor on someone else’s site. The mechanics of the link itself are the same too. An internal link has to be an a element with an href before Google will follow it, as the lesson on crawlable links explains, and a menu that works only through script events can leave a whole section unlinked as far as a crawler is concerned.

Google staff have been plain about how much weight this carries. Search Engine Journal reported in March 2022 that John Mueller, asked whether internal links still mattered on a site with breadcrumb structured data, said internal linking is one of the biggest things you can do on a website to guide Google and visitors to the pages you think are important. He added that structured data cannot replace it, and that a site should think strategically about what it cares about most and highlight that with its internal links. That is a statement about importance and not only about discovery. Which pages a site links to most, and from where, is a signal the site controls completely, which is rare in a field where most signals have to be earned from other people.

Orphan pages

A page becomes an orphan in ordinary ways. A redesign drops a section from the menu. A campaign landing page is published and promoted by email, and nobody links to it from the site. A product is removed from its category but the page stays live. An old guide falls off the end of a paginated blog archive. In each case the page still exists, but no visitor or crawler can reach it by clicking. A sitemap can get the page discovered. It passes no value and gives no anchor text, and a page that the site itself never links to is one the site is, in effect, telling Google it does not think is important.

For off-page work there are two versions of the problem, and they need opposite fixes. The first is an orphan that should rank. It receives nothing from the homepage or from any other page that has earned links, so whatever it achieves it achieves alone, and the fix is to link to it from relevant pages that have value to pass. The second is an orphan that has backlinks. This case is easy to miss, because the pages most likely to be forgotten in a redesign are old campaign pages, reports and press pages, which are the very pages journalists linked to. The value those links brought is stuck on a page that links nowhere useful, or that links only to other dead ends. The fix there is to add links from that page into the current site, or, if the page is truly obsolete, to redirect it to the nearest living equivalent.

Finding orphans is a matter of comparing lists. A crawler that follows links from the homepage produces the set of pages that can be reached by clicking. The sitemap, the analytics landing pages report and the list of linked pages in a backlink tool each produce a set of pages that exist. Any URL in the second group and not in the first is an orphan or close to one. The same crawl gives click depth, the number of clicks from the homepage to each page. A common rule of thumb is to keep important pages within three clicks, which is a habit and not a figure from Google. Matt Cutts made the underlying point in 2009 with the example of a shop that puts its products front and centre compared with one that buries them where visitors and search engines must click through many links to reach them.

Why PageRank sculpting stopped working

After nofollow was introduced in 2005, site owners noticed a side effect in how Google handled it. A page’s PageRank was divided among its followed links only. If a page had ten links and five were nofollowed, the remaining five each received a fifth. It looked like a free way to concentrate value: nofollow the links to the privacy policy, the login page and the contact form, and the product links would get a larger share. The practice was named PageRank sculpting.

On 15 June 2009 Matt Cutts, then head of Google’s webspam team, wrote on his blog that this no longer worked and had not for some time. He used a page with ten PageRank points and ten outgoing links, five of them nofollowed. Originally the five followed links would each have passed two points. Google had changed this more than a year earlier, he wrote, so that the five followed links pass one point each. The nofollowed links still take their share out of the division and pass it to nobody. Nofollowing an internal link therefore removes value from the site’s own circulation and gives nothing to the links that remain. One reason he gave for the change was that some sites trying to redirect PageRank had ended up cutting off sections that held high-quality information, such as user forums.

His advice in the same post is still the best short summary of the subject. He would let PageRank flow freely within a site, and he said the first things to attend to are making content that attracts links and choosing a site architecture that people and search engines can crawl. He allowed for a very small number of exceptions, such as links to a shopping cart or a login page. He also quoted his own earlier remark that a more effective form of sculpting is choosing which things to link to from your homepage. One thing has changed since. Google has treated nofollow as a hint since 2019, as the lesson on link attributes sets out, which makes the result of nofollowing an internal link less predictable than it was. It does not make sculpting useful again.

A working method for passing backlink value on

Nothing below is a rule from Google. It is a routine that follows from the model and from the statements above, and it is best run after each campaign and again every quarter.

1. List your pages by number of referring domains, using any backlink tool.
2. Fetch each of the top pages and check that it returns a normal page, not an error or a redirect chain.
3. For each one, pick the one or two commercial or priority pages closest to it in subject.
4. Add a link to those pages inside the main text, with an anchor that describes the target.
5. Crawl the site and list the pages with no inbound internal links, then link to or redirect each.
6. Check that priority pages are a short path from the homepage.
7. Record what you changed, so that the effect can be read against rankings later.

Step two exists because linked pages die. A report is taken down, a URL changes in a migration, and the backlinks that pointed to it now reach an error page, which passes nothing to anyone. Restoring the page or redirecting it is the cheapest link building there is, and the process is described under link reclamation. Step four leans on the reasonable surfer idea covered in link position on the page: a link in the body of an article, where a reader would plausibly click it, is likely to count for more than the same link in a footer. Imagine a small firm of accountants whose tax deadline calendar is the page that local news sites and business groups link to every January, and suppose the calendar links only to the homepage. One sentence in the calendar pointing to the self-assessment service page, with those words as the anchor, connects the most linked page on the site to the page the firm needs to rank.

This is also the reason to plan internal links before a campaign and not after it. When you build a linkable asset such as a statistics page or a free tool, decide at the outset which commercial pages it will link to and write those links into the first version. Coverage arrives in the days after launch, and a page that already passes its value on wastes none of that period.

The limits of internal linking

Internal links move value around a site and do not create it. A site with no backlinks has nothing to pass, and rearranging its links will not make up for that. The effect of any one internal link is also not something you can measure from outside, since Google publishes no figures and the tools only estimate. The model does say that a page’s value is divided among its links, so a page that links to two hundred others passes a thin slice to each, and that is an argument for a few chosen links in the text over a block of fifty. The handling of links repeated across every page of a site is a separate matter, covered in sitewide and footer links.

Anchor text deserves some restraint as well. You control internal anchors completely, and that makes it tempting to use the same exact keyword on every link to a page. Google’s guidance asks for anchors that read naturally and warns against cramming keywords into them, and the anchor text lesson explains why variety is the safer habit. The last limit is structural. Internal links survive only as long as the URLs do, and a migration that changes addresses without redirects undoes both the backlinks and the internal paths that carried their value, which is why domain migrations are planned link by link. The terms used here are defined in the glossary.