What is nofollow?

What is nofollow?

What relationship does nofollow describe?

Nofollow describes a qualification on a link rather than a quality score for the destination. It lets a publisher avoid implying the usual endorsement relationship. Google’s outbound-link guidance distinguishes this use from advertising and user-generated-content qualifications. The destination’s access and indexing policies remain separate decisions.

Primary evidence: outbound-link guidance. Accessed October 8, 2026.

How is nofollow added to an individual link?

An individual link carries the value through its anchor element’s rel attribute. The destination remains in href, and the visible text remains the anchor. Inspect all three parts when auditing implementation because a qualification, a valid destination, and a useful customer-facing label serve different functions.

The following markup is illustrative syntax using a reserved example address. It describes the relationship without asserting that the destination is a real business recommendation. A publishing system may provide the same configuration through an editor field, but the final rendered markup should still be checked.

<a href="https://example.com/reference/" rel="nofollow">
  The offer discussed in this example
</a>

The href keeps an ordinary destination available. The rel value communicates the qualification. The anchor tells the reader why the destination is relevant. A script-only button with a nofollow-like setting is not equivalent to this ordinary anchor unless the published output actually creates the supported link markup.

• Some systems automatically append values to external links.

  Check whether an editor override replaces the complete rel attribute or adds to existing values. Removing one value should not accidentally remove unrelated browser-security values that the site’s implementation requires. Review the resulting attribute rather than assuming every item in it controls search endorsement.

• Test a representative published page after the change.

  Editor previews, draft views, and final output can differ. An extension badge can help locate candidates, but the element’s actual markup is the evidence. Trace a repeated error to the template or plugin that generates the link instead of fixing the same rendered symptom manually on many pages.

What does Google’s hint model mean?

Google treats nofollow, sponsored, and ugc as hints about how links should be considered within Search. That differs from a promise that a particular qualified link is always ignored in every processing stage. Google’s announcement explaining the hint model documents the distinction.

Primary evidence: announcement explaining the hint model. Accessed October 8, 2026.

• The publisher can communicate the relationship, but cannot dictate a precise ranking calculation.

  Google may use information from the link alongside other signals. This is why a link audit should report the observed qualification without assigning an invented numerical amount of ranking benefit to it.

• A hint is also not a private access rule.

  The destination may be discovered through another ordinary link, a sitemap, or an existing known address. A business should not rely on nofollow to keep a public customer document out of search or to prevent another program from retrieving it.

• The historical announcement explains a change in handling.

  Current implementation decisions should also follow Google’s present outbound-link guidance. Using the older announcement as evidence for the concept does not justify claiming that every detail of a particular search product remains unchanged indefinitely. Retain the access date and check documentation when the workflow changes.

• Avoid the conclusion that qualified links are worthless.

  A useful reference can still help a customer verify information or reach a relevant resource. Search endorsement is one dimension of a link. Referral usefulness, editorial transparency, and the appropriateness of the destination require separate assessment.

When should sponsored be used instead?

Sponsored is the preferred relationship value for advertisements, sponsorships, or other compensated link arrangements under Google’s guidance. Nofollow remains accepted for identifying these links, but the more specific value communicates their origin more clearly. Classify the arrangement before selecting the attribute instead of judging it from the destination’s topic alone.

• Compensation can be connected to a placement even when the destination is useful.

  A supplier sponsoring a contractor’s event page creates a different relationship from an independent citation to the supplier’s installation documentation. The same destination host can therefore appear through links with different origins and appropriate qualifications.

• Do not present a paid placement as an independent editorial recommendation.

  Google treats links created primarily to manipulate ranking signals as a spam-policy concern. Its spam policies, accessed October 8, 2026, provide the relevant platform rules rather than a promise that a paid arrangement can be made safe by cosmetic wording.

• An appropriate qualification is part of publishing the relationship accurately.

  Visible copy should also explain a sponsorship where needed. Link qualification does not replace a visible sponsorship disclosure. The business should review the actual advertising context with the responsible publisher and advisers.

• When updating an existing implementation, preserve supported qualifications during the transition.

  Google’s guidance does not require removing all existing nofollow markings just because sponsored is more specific. A template-level improvement can introduce the preferred value for new compensated links while retaining coherent treatment of older placements.

When should ugc be used instead?

Ugc describes user-generated-content links, such as links contributed through comments or forum posts. Their origin differs from links selected directly by the publisher. Google’s guidance recommends this qualification for that context, while allowing publishers to consider their established moderation arrangements and the trust attached to particular contributors.

• A service website may permit customers to leave questions below an advice article.

  A link inside a submitted question belongs to that contribution. It is not automatically the business’s editorial endorsement, even when the business moderates the comment before publication. The qualification should reflect the source of the link.

• Moderation remains necessary.

  Ugc does not make irrelevant promotion or a dangerous destination acceptable. Review what the contribution adds to the customer discussion and whether the link belongs there. A meaningful question with a relevant resource differs from repeated messages inserted solely to advertise another website.

• The publisher can combine relationship values where the arrangement genuinely has more than one characteristic.

  A sponsored user contribution can differ from an ordinary customer question. Preserve a clear record of how the publishing system assigns those values instead of adding every possible attribute to every link without understanding the origin.

Can relationship values be combined?

Relationship values can be combined when they accurately describe the link. Google’s documented examples allow combinations such as user-generated and sponsored relationships. The purpose is to communicate the real arrangement, not to accumulate attributes until a tool gives the link a preferred badge or color.

An illustrative combined anchor can use a space-separated rel value. The destination is again a reserved example address rather than a verified commercial placement. Its visible label should disclose the role of the destination instead of pretending the contribution is an independent recommendation selected by the business.

<a href="https://example.com/contributor/" rel="ugc sponsored">
  Resource supplied in a sponsored contribution
</a>

A backward-compatible combination may also include nofollow where a system needs that value. The current guidance supports combinations, but the implementation should have a documented reason. Adding contradictory-looking values because different plugins compete over the same link makes future maintenance harder and can conceal the original relationship decision.

• Distinguish search-related values from other rel values.

  Some attributes affect browser behavior or opener relationships. They should not be interpreted as evidence that the publisher withheld search endorsement. Inspect the complete attribute, then classify each value by its documented role rather than treating all external-link settings as synonyms for nofollow.

• Test the final output after an editor saves the page.

  Some systems normalize, strip, or replace attributes during publishing. A correctly entered field is not enough if the rendered anchor loses the qualification. Repeated omissions usually require fixing the editor or template behavior instead of asking every writer to paste custom markup indefinitely.

• Inspect links that pass through a tracking redirect separately.

  The rel value belongs to the anchor on the source page; it is not a response header carried through each redirect. Check the visible destination, actual href and resolved destination together. A redirect can change where the visitor lands without changing the relationship recorded on the original link.

• For an illustrative sponsored supplier link, record which placement was compensated and which publisher controls its HTML.

  If another system generates the outbound URL, inspect its output rather than assuming that its tracking label supplies the required relationship value. This separates the commercial arrangement, link markup and destination behavior into evidence that can actually be checked.

How does link-level nofollow differ from page-level nofollow?

Link-level nofollow applies to a particular anchor relationship. A page-level robots nofollow instruction concerns links on the page under the documented processing rules. The difference matters during diagnosis: changing one anchor may leave a broader instruction in effect, while removing the broad rule may leave deliberately qualified anchors unchanged.

Why is nofollow not an indexing exclusion?

Nofollow qualifies a link relationship; it does not require exclusion of the destination from search. Google can know and process the destination through other sources. The appropriate indexing control depends on the destination owner’s intended policy and whether the supporting crawler can reach and read the relevant instruction.

• A business cannot reliably hide its own public customer route by placing nofollow on the one visible link leading to it.

  Another page may link to the same address. A sitemap may list it. Google may already know it. The destination remains publicly retrievable unless genuine access control says otherwise.

• Noindex is the separate supported page-level exclusion mechanism.

  It must be readable on the resource for Google to process it. A robots.txt restriction concerns crawling instead. Choosing the correct control requires defining whether the goal is link qualification, search exclusion, or restricted access to information.

• The publisher also does not control an external destination’s policy through its outgoing attribute.

  If a referenced page contains sensitive information, the destination owner needs to address that resource. Removing or qualifying one reference does not erase the address from other sources or prevent direct visits.

• Record the intended outcome before implementing the link change.

  If the goal is to describe a compensated relationship, use the appropriate qualification. If the goal is to remove a public page from supported indexing, investigate the page’s instruction. A clear requirement avoids a superficially successful link edit that never addressed the real problem.

Should ordinary internal links use nofollow?

Ordinary internal links should be reviewed according to their navigational purpose rather than treated as a ranking-allocation mechanism. A business needs customers and crawlers to reach relevant services and advice through coherent routes. Broadly qualifying those routes does not provide a reliable way to direct a precisely measured amount of ranking benefit elsewhere.

• An internal service link introduces another part of the same website’s explanation.

  It normally has a different role from an external paid placement or unreviewed comment. Review the source page, destination, and reason for the qualification. An inherited plugin rule that marks every link may obscure the intended internal architecture.

• Internal linking focuses on meaningful routes and context.

  A useful water-heater guide can be introduced from the relevant service explanation through an ordinary descriptive anchor. The business should not make that route less clear because a checklist suggests that every link spends a measurable unit of a hidden budget.

• Noindex or utility-page decisions belong at their appropriate boundaries.

  A confirmation route may not need an ordinary browsing introduction at all. If it remains publicly linked within an operational workflow, its indexing policy still requires its own instruction. Nofollow is not a substitute for defining that policy.

How should backlinks marked nofollow be interpreted?

A qualified incoming link is an observation about the publisher’s markup and relationship statement. It is not a complete measure of its business value or Google’s treatment. The destination owner can inspect the observed anchor and qualification, but should avoid assigning invented ranking-credit numbers to a link from an external website.

• A relevant local resource may send customers through a qualified link.

  A widely distributed spam mention may send nobody despite an ordinary rel state. Compare the link’s context, destination relevance, and actual referral evidence where available. The attribute alone does not establish either a useful referral or an unsuitable placement.

• Backlinks need source and context assessment.

  A report label can help organize them, but cannot establish authenticity of editorial selection or audience fit. A crawler may also miss links behind unsupported rendering or a login. Keep the collection method and observed period with any comparison.

• Do not request removal of every nofollow value as an automatic outreach goal.

  The publisher may have a valid compensated or contribution policy. Asking them to misrepresent that relationship creates a different problem. If a link is genuinely broken or the anchor misdescribes the service, address that concrete issue instead.

• A placement decision should be defensible without a promised ranking transfer.

  Useful coverage, accurate business information, and a relevant customer route can matter independently. A business should reject offers whose main justification is an assured benefit from changing one attribute, because the available evidence does not support such precision.

How should anchors and destinations be checked together?

A link audit should inspect the relationship, destination, and visible anchor as separate but connected fields. The rel value describes the qualification. The href identifies where the visitor goes. The anchor explains what the link offers. A correct attribute cannot compensate for an unrelated target or misleading wording.

• Anchor text should help a customer predict the destination.

  A reference to a supplier’s maintenance instructions should not lead to a generic promotional homepage without explanation. A sponsorship link should not be labeled as an independent technical source. The editorial wording needs to match the real role of the destination.

• Request the destination and follow any relevant redirects.

  A link can become outdated after a publisher changes routes. An ordinary-looking address may lead to a parked domain or unrelated offer. Qualification does not eliminate the need for destination review, especially when old articles continue exposing the link to customers.

• Keep links crawlable where discovery is intended.

  Google’s crawlable-link guidance, accessed October 8, 2026, recommends normal anchors with usable href values. Adding a rel value to a visual component does not establish that the component exposes an actual destination in supported markup.

• Prioritize links by customer relevance and exposure.

  A broken primary-document reference in a service decision guide can matter more than a dormant qualified link in an old event announcement. Classification should help choose the repair, not convert every link into an equal point on a mechanical checklist.

An illustrative decision: a sponsored resource and a source citation