Should you use nofollow on internal links?

Updated in 2026: Advice about “PageRank sculpting” led many sites to add nofollow to internal links. That confuses several separate controls and rarely improves a site.

Usually, you should not use nofollow on internal links. Normal links help people and crawlers discover your pages. nofollow is not a reliable way to prevent indexing, manage canonical URLs or sculpt authority; use the control designed for the actual problem.

Choose the control by intent

• To prevent a page appearing in search, allow crawling and use a supported noindex directive.
• To prevent crawling of a path, use robots.txt carefully. Blocked URLs can still be discovered and may appear without content.
• To choose among duplicate URLs, use redirects, canonical links and consistent internal linking.
• To protect private content, require authentication. Search directives are not access control.
• To mark a paid or user-generated external relationship, use rel="sponsored" or rel="ugc". Use nofollow where the other values do not fit.

Google’s current link qualification guidance says ordinary links need no rel attribute. It also notes that qualified links may still be discovered through other routes.

When internal nofollow can be reasonable

There are unusual cases where you genuinely do not want Google to associate with or crawl a linked destination and another control is not available. Treat that as an exception and document why it exists.

Do not add nofollow to login, basket or filter links as a substitute for designing crawl behaviour. Do not rely on it to preserve a fixed quantity of “link equity”.

Start with the user journey and the indexing requirement, then apply the narrowest correct control. If the reason is merely “an audit tool told us to”, the site probably needs a clearer diagnosis.