内部链接

Any link whose destination sits on the same domain counts: the menu, the footer, 面包屑, a related-posts block, a link mid-sentence, a logo pointing home, an image wrapped in a link, a button. All of them divide up the value the page has to pass on.

What does not count: links to another domain — those are external links — links that only fire through JavaScript with no 链接 to follow, and anything behind a login. A crawler can only follow what stands in the HTML as a real anchor with a real address. A site whose navigation is a JavaScript menu without a fallback has, from a crawler’s point of view, almost no internal links at all.

• Discovery. A crawler arrives at the homepage and follows links. If nothing points at a page, there is no route to it — and 点击深度 decides how often the ones that are reachable get revisited.
• Distribution. Whatever standing a domain has flows along its internal links. This is what people mean by link equity.
• Classification. The words in the link and the sentence around it tell a search engine what the target is about. Twenty pages pointing at one page with related wording is a strong, consistent signal.

The first two are mechanical. The third is where the craft is, and it is called 锚文本.

重要的区别不在于 哪里 a link sits but whether a human decided on it. Navigation links are structural: every page gets them, so they say very little about any individual page. A link inside a sentence is editorial — somebody decided that this passage was a good reason to send a reader to that page. Search engines weigh the second kind far more heavily, and so do readers.

This is why a long article with a full menu, a full footer and not one link in the prose is the normal case on the web, and also the most common thing worth fixing.