Site SEO AI AuditInternet Solutions ürünü

Link rel Attributes Explained: What Each One Does

12 Kasım 20268 dk okumaTeknik SEO
Link rel Attributes Explained: What Each One Does

Short answer: The rel attribute describes the relationship between the current page and the resource it links to. On links in the page body, nofollow, sponsored and ugc tell search engines how to treat the link, while noopener and noreferrer protect visitors and privacy. In the head of the page, rel appears on link elements that declare the canonical URL, alternate language versions, feeds, icons and performance hints such as preload and preconnect. Use only the values you need, and test them.

The word rel is short for relationship, and it appears in many places in HTML. A single attribute name with dozens of possible values can be confusing, especially because some values are for search engines, some for browsers and some for security. People often add rel values copied from a plugin or a tutorial without knowing what they do, and some combinations cancel each other out or are simply ignored. This guide groups the most common values by purpose, explains what each one does and warns about the usual mistakes. For a broader look at link handling, read the guide to nofollow, sponsored and UGC links.

How the rel attribute works

Understanding the basic syntax saves a lot of confusion later, because the same word appears in very different roles.

The attribute can appear on anchor elements, which create the clickable links in the body, and on link elements, which sit in the head of the document and connect the page to other resources. Several values can be combined, separated by spaces, for example rel=”nofollow noopener”. Browsers and crawlers read the ones they understand and ignore the rest.

Because the same attribute does very different jobs, it helps to ask who the value is meant for: a search engine deciding how to treat a link, a browser deciding how to open a page or load a resource, or a tool reading metadata. The sections below follow that division.

Values that tell search engines about a link

The nofollow value asks search engines not to follow the link or to pass ranking signals through it, and search engines treat it as a hint rather than a command. Two more specific values were added later. The sponsored value marks links that are paid or part of an advertising arrangement, and the ugc value marks links in user-generated content such as comments and forum posts. You may combine them, for example rel=”sponsored nofollow”.

Value Use it for
nofollow Links you do not want to vouch for, when no more specific value fits
sponsored Advertisements, paid placements and sponsorship links
ugc Links added by users in comments, reviews and forums

Do not use these values on ordinary internal links to hide pages from search; they are not a reliable way to control indexing. Use noindex or proper access rules for that. And never use sponsored links without marking them: undisclosed paid links break search spam policies.

Values that protect visitors: noopener and noreferrer

When a link opens a page in a new tab, the new page can, in older browsers, get limited access to the page that opened it. The noopener value prevents this, and modern browsers apply it automatically to links with target=”_blank”, but adding it is still a common safeguard. The noreferrer value stops the browser from sending the address of your page to the destination site as a referrer. That protects privacy but removes the information from the destination’s analytics, so use it only when you want that effect.

These values have no effect on rankings. They matter for security and privacy, and they appear in many content management systems by default for external links that open in a new tab.

Values in the head: canonical and alternate

In the head of the document, rel works on link elements that describe the page itself rather than a clickable path for the visitor.

The link rel=”canonical” element names the preferred URL for a page whose content is available at several addresses. It is a hint, and search engines may choose another address if signals disagree, as discussed in the guide to canonical tags. Each page should have at most one canonical, and it should point to an address that returns a normal success status.

The link rel=”alternate” element has several uses. With the hreflang attribute, it lists language and country versions of the same page. With a type attribute, it can point to an RSS or Atom feed. In older setups for separate mobile sites, it pointed to the mobile version. The most important point is that alternate versions should reference each other, so that each language version lists all the others and itself.

Pagination values: prev and next

Earlier, link rel=”prev” and “next” were used to show that pages belong to a sequence, such as page two and page three of a category. Google stated that it no longer uses them for indexing. Other tools and browsers may still read them, so they are harmless if already present, but you should not rely on them. Make each page in a sequence useful on its own, link between pages with normal anchors, and keep canonical tags self-referencing. The advice in the article on pagination goes deeper.

Performance hints: preload, preconnect and friends

These values belong to speed work rather than to link policy, yet they use the same attribute and often sit in the same files, so they are easy to mix up.

Several rel values tell the browser to start work early. Preload asks for a critical resource, such as a font or the main image, as soon as possible. Preconnect opens a network connection to another domain you will use soon. Dns-prefetch resolves a domain name in advance, and prefetch fetches a resource that will probably be needed on the next page.

These hints can improve loading, but overuse backfires: preloading too many files makes them compete for bandwidth, and preconnecting to many domains wastes effort. Use them for a few critical resources and measure the effect. The article on resource hints shows how.

Other values you may meet

The standard defines many more values than most sites ever need, and most of them are harmless to leave in place.

Stylesheet links to CSS files, and icon links to the favicon. Manifest points to a web app manifest. Author and me describe the relationship to a person or a profile, and search engines do not use them as strong signals today. Help, license and search also exist in the HTML standard, with limited practical effect on SEO. If you see a value you do not recognise, look it up before deleting it, because some are used by browsers, readers or plugins.

Documentation for each value is published in the HTML standard and in the help pages of the search engines. Avoid inventing your own values. A made-up rel value is ignored, and it can mislead colleagues who assume it does something.

Common mistakes with rel

The first is nofollowing all external links by default, a habit from years ago. It cuts you off from the normal web of references and does little for ranking. The second is the opposite: leaving paid links unmarked, which risks a penalty. The third is putting canonical tags inside the body of the page or listing two different canonicals, which makes search engines ignore both.

A fourth mistake is pointing canonical to a redirected, blocked or noindex page, so the signal contradicts itself. A fifth is adding hreflang alternates without return links, which means the annotations are ignored. A sixth is overusing preload: a page that preloads ten files may load slower than one that preloads two. Finally, many sites let several plugins print the same tags, so a page ends up with duplicate canonicals or repeated alternates. View the page source, search for rel=, and count what you find.

When you fix a problem, write down which template or setting caused it and test two or three different page types. Rules in content systems often differ between posts, pages, categories and archives, and a fix that works on one may not reach the others.

Check your own links

Review the rel values on your site in two places: external links in the body, where nofollow, sponsored and ugc should match the real relationship, and the head, where canonical, alternate and hints should be correct and not duplicated. A crawl with Site SEO AI Audit checks links, canonicals, language annotations and more across seven areas, and the first audit of up to 200 pages is free.

When you find a problem, fix the template or plugin setting that creates it, not just the single page. Most wrong rel values come from one rule applied to many pages.

Related reading

The bottom line

Use nofollow, sponsored and ugc to describe links honestly, noopener for safety, and canonical, alternate and hints in the head only where they are correct.

SSS

Does nofollow stop a page from being indexed?

No. It is a hint about the link. To keep a page out of results, use noindex or restrict access.

Should all external links be nofollow?

No. Link normally to sources you trust, and use sponsored or ugc where they describe the relationship.

Do I need noopener on every external link?

Modern browsers add it for links that open in a new tab, but including it is harmless and protects older ones.

Are prev and next still useful?

Google no longer uses them for indexing. Make paginated pages work through normal links and self-referencing canonicals.

Can I combine rel values?

Yes, separate them with spaces, for example sponsored nofollow noopener.

#Backlinks#Canonical tags#On-page SEO#Technical SEO
Kendi web sitenizi denetleyin.Google web sitenizi görüyor mu?
Ücretsiz başla

Blogdan daha fazlası

Tüm makaleler →
Internet Solutions

Ekibimizden diğer ürünler

Internet Solutions tarafından geliştirildi. Diğer ürünlerimizi de deneyin — her biri size farklı bir şekilde zaman kazandırır.

internet-solutions.net ↗