An orphan page is a page on your site that no other page links to. It exists, it is indexable, and as far as your internal link graph is concerned it may as well not be there. On most sites with a few years of publishing history, somewhere between 5% and 20% of pages are orphaned or nearly so.
Why they fail to rank
Internal links do two jobs, and orphans miss both.
Discovery. Crawlers find pages by following links. An orphan can still be discovered through a sitemap, but sitemap-only URLs are crawled less frequently and treated with less confidence than pages the site itself points to.
Equity. Internal links pass authority. A page with zero internal links inherits none of your domain’s accumulated authority and starts from scratch every time — which is why orphaned pages often sit at position 40 while similar linked pages on the same site rank on page one.
There is a third, quieter cost: an orphan is a page you are paying to host and maintain that returns nothing.
Where orphans come from
Almost all of them are created by one of these:
- CMS collection pages. Items reachable only through a paginated listing. Page one gets linked; item 340 does not.
- Migrations. New navigation drops links that used to exist. The pages survive, the links do not — which is why this belongs on every migration checklist.
- Campaign landing pages. Built for a paid campaign, never linked from the site, left indexable after the campaign ends.
- Old blog archives. Posts that fell off the paginated index and were never linked contextually from anything newer.
- Filtered and faceted URLs. Generated by parameters, indexable, linked from nowhere stable.
How to find them
You need two datasets, because neither alone is sufficient.
1. Crawl your site. Start from the homepage and follow links only. This produces the set of pages that are actually reachable.
2. Get the complete list of pages that exist. Combine your XML sitemap, a CMS export of all published URLs, Search Console’s indexed pages, and analytics landing pages for the last twelve months.
Orphans are set two minus set one. Anything in the “exists” list that your link-following crawl never reached is an orphan.
Most crawling tools do this natively — Screaming Frog can ingest sitemap and Search Console data and flag the difference directly. Do not skip the analytics source: pages that still receive traffic despite being orphaned are your highest-priority fixes, because they are already proving demand with no support at all.
Also look for near-orphans: pages with one or two internal links, usually from a paginated archive. They are not technically orphaned but they behave almost identically, and there are normally far more of them.
How to fix them
Do not just add links. Decide what each page is for first.
Keep and link properly. If the page has value, give it contextual links from relevant existing content — ideally from pages that already have authority. Two or three genuinely relevant contextual links beat twenty from a footer block. Anchor text should describe the destination; the internal linking strategy guide covers how to choose it.
Merge. If the page overlaps heavily with something better, consolidate the content into the stronger page and 301 the orphan to it. This is usually the right call for thin archive posts.
Remove. Expired campaign pages and duplicates should 410, or be noindexed if they still serve a purpose for users.
Fix the template, not the page. If your CMS orphans collection items by design, the durable fix is structural: hub pages that link to every item, related-item modules, breadcrumbs, and archive pages that are actually reachable. Otherwise you will re-orphan everything you publish next year.
Prevention
Add an orphan check to your publishing process. Every new page should get at least two contextual internal links from existing content before it goes live — a two-minute step at publication that saves the audit later.
If you want the whole graph mapped and a specific list of links to place, that is exactly what an internal linking audit produces: source URL, destination URL, anchor text, ready to implement.
