孤立页面
An orphan page is a page that no other page on the same site links to. It loads perfectly and may well sit in the XML 站点地图, but nothing following links can reach it.
That is the whole criterion: zero inbound internal links. Almost every site older than a year has some, and they are usually created by tidying up rather than by neglect.
Three things it is not
- Not a 404. The page exists and works. That is exactly what makes it easy to miss — nothing is broken.
- Not a noindexed page. Excluding a page from the index is a decision; orphaning is an accident.
- Not saved by the sitemap. A sitemap tells a crawler a URL 存在. Internal links tell it the URL 事项, and roughly how much.
One near-miss is worth naming: a page reachable only through page seven of a paginated archive is technically linked, but from a page that itself sits deep and carries almost nothing. Practitioners call these 近乎孤儿 and treat them the same way.
孤立页面是如何产生的
Nobody creates one on purpose. They accumulate through ordinary maintenance: a post drops off the blog listing, a campaign ends but its landing page stays live, a menu item is removed to simplify the header, content is migrated without the links between it, a category is restructured and its children are cut loose.
The common thread is that the page was linked once and the 链接 disappeared through a change nobody connected to it. Every simplification of a navigation buries something.
Finding them
The method is a subtraction: everything that exists, minus everything a crawler reaches by following links. What remains is orphaned.
Crawl from the homepage with Screaming Frog or Sitebulb, feed the sitemap in as a second source, and both tools flag the difference directly — in Screaming Frog under Crawl Analysis as 孤立网址. Search Console alone will not answer it: its Links report shows the most-linked targets, not the unlinked ones, and absence from a top list is not proof of absence.
What it costs
Three separate costs, and the third is the one that stings.
爬取。. Pages nothing links to are visited rarely, so changes to them are noticed late.
权威。. Whatever standing the domain has flows along internal links. A page receiving none inherits none and competes on its own from a standing start.
这项工作已经付过钱了。. Somebody researched, wrote and edited that page. It is still good. It simply cannot be found through the mechanism most of the site relies on. Orphans are usually the cheapest traffic available precisely because the expensive part is already done.
链接、合并或删除
Do not link everything back reflexively. Each orphan gets one of three verdicts, and one question decides it: 你今天要发布这个页面吗？
- 将其连接起来 — if yes. Find two or three pages whose subject genuinely leads there and link from inside the text, not from a footer list.
- 合并 — if the answer is “not on its own, but the content is worth keeping”. Fold it into a stronger page on the same topic and redirect the old URL there.
- 删除它 — if no. Delete it and let it return 410, or redirect it if it has inbound links or lingering traffic.
Work through them by traffic, highest first. An orphan with impressions is already ranking with one hand tied.
Not producing new ones
Finding orphans is a cleanup. Not creating them is a habit, and it comes down to three things.
链接随出版物发布，不迟于此。. 当页面上线时，立刻决定好哪两三个现有页面应该指向它——并编辑它们。“我以后再加链接”正是积压工作开始的方式。.
将结构变化视为链接变化。. Removing a menu item, retiring a category, unpublishing a hub: each one cuts links. Ask what loses its route before making the change, not after.
按计划重新抓取。. Quarterly is enough for most sites, takes twenty minutes, and catches the drift before it becomes a project.
English 
Deutsch 
Español 
简体中文 
हिन्दी 
Français 
العربية 
Русский 
Português