Fixing Broken Internal Links: A Guide for Webmasters and Small Business Owners

A broken internal link is a link on your own website that no longer leads to the page or resource a visitor reasonably expects. It may return a 404, pass through a loop, open an obsolete document or lead to a generic homepage instead of the promised service. The consequence can be simple frustration or a lost enquiry at an important decision point. Fixing internal links therefore requires more than a site crawler and a spreadsheet of error codes. You must identify what the link was meant to accomplish, choose the appropriate destination or removal, and verify that the updated public page still supports its reader's task.

Know which links count as internal before auditing them

Internal links connect resources within the website or its relevant owned site structure. They appear in navigation menus, article paragraphs, category cards, breadcrumb trails, downloadable resource areas and calls to action. A linked company logo can be as important as a reference in a long guide. To audit reliably, establish which hostnames and subdomains belong to the relevant property and whether other company sites are intentionally treated as separate destinations. A crawler configured to inspect only one narrow path can miss significant errors elsewhere.

Record both the source URL containing the link and its destination. A report listing only a broken target does not tell the editor where to fix it. Record the anchor wording or component label because it expresses the promise made to the visitor. A link called 'book a consultation' deserves a different priority from a broken citation in a retired article. Where the same bad link appears in a shared footer, distinguish the one underlying template defect from the many pages exposing it.

Use a crawl to discover candidates, not to declare every error proven

A site crawler can collect href values and identify common response patterns, including missing pages, redirects and errors. Choose a crawl size and rate appropriate to the site, and obey applicable access restrictions. Capture the timestamp, requested URL, observed response and source page. Some automated tools cannot follow JavaScript-only navigation correctly; others may encounter anti-bot protections or transient responses that do not affect ordinary visitors. Treat those as candidates requiring interpretation.

Test the important failures in a browser or by inspecting the actual response. Look for a not-found message hidden inside an HTTP 200 page, a redirect that ends on an unrelated subject, a response that depends on being signed in, or a download that presents the wrong document. A green result from a crawler proves little if the destination is technically available but semantically wrong. The acceptance question is whether a person following the link gets the content or action that the label promised.

Distinguish 404, 410, redirects and server faults

HTTP 404 generally means a requested resource was not found. HTTP 410 can indicate that it is intentionally gone. A 301 or 308 represents a permanent redirect in appropriate settings, while 302 and 307 express temporary redirection semantics. Server errors in the 5xx range indicate a different class of problem. These definitions matter because the same fix cannot address an obsolete article, a mistakenly typed path and an unavailable application server.

Check the destination's history before acting. If a guide moved to an equivalent updated URL, the link can be changed directly and a suitable redirect may help bookmarks and external references. If the article was retired without replacement, update or remove the link rather than forcing an unrelated destination. If server faults are intermittent, investigate hosting, access rules and logs with the responsible technical team. Do not create a maze of redirects simply to make a report show fewer 404s.

Fix the source link first when the correct URL is known

A common error is to rely on a permanent redirect instead of updating the outdated link embedded in the website. Redirects can be useful for previously shared URLs, but your own navigational and editorial links should generally point directly to the preferred destination. Open the source page in the correct content editor or template, replace the broken href and confirm the anchor still accurately describes the target. A link pointing to a new generic service page may require a change of wording as well.

Where a link appears on dozens of pages, identify the generating component before making individual edits. A shared navigation item, CMS block or product-card template may control all occurrences. A one-time template fix can solve the root cause, but it also has a wider regression risk. Test representative page types after deployment. Maintain a changelog showing who made the correction and the exact final destination so that later editors do not restore the old address from copied content.

Handle retired pages without misleading visitors

Some removed content genuinely has no successor. An old event announcement, an expired product or superseded policy might not warrant restoration. Review whether the original link still has editorial value. Remove the reference when it no longer supports the page, or replace it with a current authoritative resource if one exists. If the surrounding sentence depends on the removed document as evidence, revise that sentence rather than leaving an unsupported claim.

Do not automatically redirect every missing page to the homepage. Somebody following a precise troubleshooting link expects technical guidance, not a general company introduction. If the old URL continues to attract visitors, an honest, helpful not-found page can explain that the material is unavailable and offer relevant navigation. The error page should not falsely return a successful status while hiding the underlying absence. Preserve useful alternatives without pretending they are identical to the retired content.

Prioritise customer journeys over raw error totals

A crawler may count hundreds of failures from one template while a single broken enquiry link appears once and causes greater practical harm. Rank issues by the importance of the affected journey, how many visitors are exposed, the ability to recover and the effort required to fix the source. Broken contact and checkout actions, essential policy links and primary navigation usually deserve careful early attention. Minor historical article references may be scheduled according to their relevance and readership.

For an agency handover, group findings into blocking, high-impact and maintenance categories. State why each category matters. If a broken link is visible only in an archived page with little traffic, that does not mean it should be ignored forever; it means the business can make a deliberate scheduling choice. An automated severity label cannot replace knowledge of what the website is built to accomplish.

Check internal links as part of indexing and site organisation

Important crawlable links can help search engines discover public pages, while meaningful anchor text provides readers and systems with context. A page available only through a sitemap can be harder for ordinary visitors to find. When repairing broken links, ask whether important pages have enough natural navigation from related sections. Add relevant connections without turning every paragraph into a keyword-rich linking exercise.

Google's crawlable-links documentation explains the importance of standard anchor elements and understandable link text. Follow it alongside the site's actual information architecture. Do not claim that changing a few anchors guarantees search positions. Internal links support discovery and navigation, while indexing and ranking remain separate search-engine decisions.

Retest redirects, canonical destinations and mobile behaviour

After updating links, test the public source page, activate the destination and confirm it fulfils the stated purpose. Check that the final URL is consistent with the site's preferred canonical and that the path does not loop through unnecessary redirects. On mobile, ensure the link is visible and can be tapped without being obscured by overlays. With a keyboard, confirm that focus reaches the link and its destination makes sense when read in context.

A successful HTTP response alone cannot prove an accessible, useful experience. A PDF link should open the intended document, a form link should lead to the right form, and a button should perform the advertised action. Recheck sitewide templates after changes to menu structures, URL routing or domain migrations. Repeat the crawl where justified but do not call the project complete unless the significant broken journeys have been directly tested.

Prevent recurring breakage during publishing and migrations

Introduce a small editorial check for newly published content: open important references, verify internal destinations and use descriptive anchor text. During migrations, preserve a URL inventory mapping genuinely equivalent old and new resources, confirm redirects and update navigation directly. When articles are retired, consider incoming links before deleting the source material. These habits reduce preventable faults more effectively than an occasional emergency crawl.

Content teams should know whether link destinations are controlled by article fields, shared CMS blocks or application routes. Developers should provide a reliable way to test changes before release. Keep a short owner and verification record so later teams can reconstruct what happened. An unresolved failure should remain openly tracked rather than disappearing because a report was edited. This matters particularly when a source citation carries legal, technical or customer-service significance.

Measure the repair by reader outcomes, not an invented ranking lift

A complete repair means that an intended source link now reaches the correct destination, the public page renders it, and the reader can use the action without confusion. Record the before-and-after URLs, link wording, observed responses and test date. Search Console may eventually report related crawl or indexing observations, but it does not provide a guarantee that a particular link fix will raise rankings. Similarly, a drop in crawler error count cannot by itself prove an increase in sales.

Focus on the understandable result: fewer dead ends, more dependable navigation and a site that is easier to maintain. For primary references see Google's HTTP and network errors guide, Google's redirect guidance, crawlable link guidance and W3C guidance on link purpose. These describe technical and accessibility expectations, not ranking guarantees.