A customer lands on your homepage, clicks a product guide, and gets a 404 page. They try once more, find another dead button, and leave. Your team may not notice the problem until a campaign underperforms or a sales representative reports that prospects can't reach an important page.
A broken link isn't just untidy HTML. It can interrupt navigation, weaken confidence, waste crawler attention, and expose a deeper maintenance problem caused by migrations, deleted content, changed URL structures, or third-party pages disappearing. The scale is easy to underestimate. A 2024 Pew Research Center analysis of link rot found that 38% of webpages that existed in 2013 were no longer available by 2023, while 54% of Wikipedia pages had at least one dead reference link.
The reliable answer to how to fix a broken link is not “add a redirect” every time. You need to identify what failed, decide whether the destination still has a useful successor, repair the source and destination appropriately, then test the result. The six steps below give marketing managers a practical workflow, whether you're also comparing mattress marketing options or managing a large publishing site.
Table of Contents
Introduction to Broken Links
A broken link is a hyperlink that leads to an unavailable, incorrect, or technically failing destination. Internal links break when a page is renamed, deleted, or moved without a matching redirect. External links fail when another organization changes its site, removes a resource, lets a domain expire, or introduces a server problem.
The distinction matters because the fix depends on ownership. You can edit an internal link directly, while an outbound link usually requires a replacement source or removal. A 404 may indicate that a page has moved, but it may also mean the address contains a typo. A temporary 5xx error shouldn't be treated like a permanently deleted page.
Practical rule: Don't repair the status code before understanding the content relationship between the old URL and the intended destination.
Link rot is ongoing rather than exceptional. The Pew analysis cited above also found broken links across news and government webpages, which shows that even heavily maintained and authoritative sites need recurring checks. A one-time cleanup can remove today's errors, but it won't protect a site after the next redesign, content consolidation, product retirement, or CMS migration.
A sound repair process has six parts: find broken URLs, validate the failure, diagnose the cause, choose a redirect or direct update, apply the fix in your CMS, and re-crawl the result. That final step turns a plausible edit into a verified repair.
Detecting Broken Links Across Your Site
Start with a crawl, not a random tour of your website. Tools such as Screaming Frog SEO Spider, Ahrefs Site Audit, and similar crawlers follow links from your pages and record the response returned by each destination. Configure the crawl to include internal links and, where appropriate, external links. Export the results so the team can work from a source-page list rather than a loose collection of browser tabs.
Build a useful error list
Filter the crawl by response class and separate the findings into actionable groups:
- 404 Not Found: The server can't find the requested resource. Investigate whether the page moved, was deleted, or was linked with an incorrect path.
- 410 Gone: The server indicates that the resource has been intentionally removed. Look for a relevant replacement before deciding whether the source link should disappear.
- 5xx responses: The server or an upstream service failed. Confirm whether the issue is temporary before changing content or creating redirects.
- Redirect responses: Record the entire path, not just the first response. A link that eventually resolves may still pass through unnecessary hops or a loop.
The crawl report should include the source page, anchor text, broken destination, response code, and redirect target when one exists. Sort first by business importance. A dead link on a homepage, pricing page, product page, or heavily visited guide deserves attention before an obscure archive page.
The scale of the problem supports routine crawling. An Ahrefs study covering 2,062,173 websites found that 66.5% of links pointing to those sites since January 2013 had rotted, and reported that roughly one link in every 200 breaks each week. Those figures don't mean every site has the same risk, but they do show why an export from last year can't serve as a permanent clean bill of health.
Validate before editing
Automated crawlers produce candidates, not final judgments. Open a sample of flagged URLs in a browser, test the destination more than once when a server error appears, and inspect the response in browser developer tools. You can also use the W3C link checker for targeted validation.
Check the canonical destination, protocol, hostname, capitalization, and trailing slash. A crawler may flag a URL because a site blocks automated requests, while a real browser receives the page normally. Conversely, a browser may display a custom error page even though the server returns a failure code that still needs attention.
If your team is also reviewing pages that search engines have crawled but not indexed, keep that diagnostic separate from link repair. The workflow described in this guide to fixing “crawled, currently not indexed” pages addresses indexation, while the present queue concerns links and their HTTP responses.
Create a working sheet with one row per broken destination and columns for source URL, link type, status, likely cause, replacement URL, chosen action, owner, and verification date. That small amount of structure prevents two common failures, repairing the same URL repeatedly and losing track of whether the source page was ever updated.
Identifying Root Causes for Broken Links
The same visible symptom can come from very different failures. Treat the response code as evidence, not a diagnosis.
Read the failure in context
A 404 often follows a URL change. For example, an old link to /services/seo-audit may fail after the page moves to /technical-seo-audit. It can also result from a misspelled slug, an incorrect file extension, a deleted product, or a link copied with an extra character.
A 410 generally communicates deliberate removal. That doesn't automatically mean you should redirect it. If the removed page has a direct successor, a redirect may help visitors. If it represented a discontinued offer with no equivalent, removing the source link is more honest than sending users to an unrelated page.
A 5xx response points toward the server, hosting layer, application, or upstream dependency. Test it again, review server and application logs, and check whether the problem affects one URL or an entire section. Redirecting around a temporary outage can hide the underlying incident and create an inaccurate destination.
Trace migrations and redirect paths
URL restructuring is a frequent root cause. During a migration, teams often update navigation but overlook old blog references, downloadable assets, image links, XML sitemaps, or campaign landing pages. Search the CMS database or repository for the old path, then inspect every source page that still uses it.
Third-party link rot requires a different investigation. The external site may have moved the resource, changed its protocol, retired the page, or become unavailable altogether. Look for a current page from the same publisher, then assess whether it supports the same claim and user intent. If no relevant substitute exists, remove the outbound link and revise the surrounding copy so it doesn't promise a resource that no longer exists.
A technical guide to broken-link testing and HTTP handling emphasizes that 404, 410, and 5xx responses require different handling, and that redirect chains longer than two hops create performance and link-equity risks. The practical lesson is simple: record the complete path from source to final destination, not merely the first error you see.
Choosing Between Redirects and Link Updates
A redirect repairs the old address for visitors and external references. A direct link update repairs the page that contains the outdated hyperlink. Neither option is universally superior.
Use a single-hop 301 redirect when the content has permanently moved and the old URL may still receive visits or inbound links. A product page that moved to a new slug, or an article transferred during a site restructure, fits this pattern. The old address remains useful as an entry point, so preserving it has a clear user and maintenance benefit.
Update the source link directly when the problem is a typo, when the replacement URL is already known, or when you control many internal references to the same destination. Direct links give browsers and crawlers the shortest route and reduce the chance that a future redirect chain will accumulate.
Remove the link when the destination has no relevant replacement. Don't redirect an obsolete webinar, discontinued product, or unrelated article to the homepage merely to eliminate a 404. That creates a technically successful response but a poor user experience.
| Situation |
Preferred action |
Reason |
| Content moved permanently |
Add a single-hop 301, then update important internal links |
Supports old references while keeping current navigation direct |
| Typo or incorrect path |
Edit the source link |
The destination was never meant to be different |
| External resource replaced |
Update the outbound link |
You don't control the external domain |
| No relevant successor |
Remove the link or return an intentional 410 |
Avoids sending visitors to misleading content |
The recommended sequence is to crawl, verify, map each broken destination to its source pages, then apply the remediation flow described in the published link-rot guidance. Use a redirect only for a permanent move. Otherwise, update or remove the link.
The safest CMS fix is the one that changes the correct layer. Edit the link when the source page is wrong. Add a redirect when the old address must continue working for users and external references. Publish, clear relevant caches, and test outside the administrator session.
WordPress
For a direct update, open the post, page, menu, button, or widget containing the link. Select the linked text or element, replace the URL, save the draft, and publish. For a moved URL, use a redirect manager such as the Redirection plugin, add the old relative path as the source, enter the new path as the target, and choose a permanent redirect.
Then check the permalink settings under Settings > Permalinks if WordPress has recently changed URL behavior. Don't repeatedly save permalink settings as a substitute for a planned redirect map, and don't install several redirect plugins that may process the same request differently. Teams publishing frequently may also benefit from a documented WordPress publishing workflow that records URL changes before publication.
Shopify
In Shopify Admin, open Online Store > Navigation, then locate View URL redirects. Choose Create URL redirect, enter the old path in the “Redirect from” field, enter the new path in “Redirect to,” and save. Use a direct internal link edit when menus, collections, product descriptions, or theme components still point to the old address.
For links embedded in a theme, open Online Store > Themes, choose the relevant theme, and use Edit code only when the URL is hard-coded in a Liquid template. Search the theme files for the old path, replace it with the canonical destination, and test both the storefront and responsive navigation. Avoid redirecting every retired product to the homepage, since customers may arrive expecting a specific item.
Wix
To edit an internal link, open the Wix Editor, select the text, button, image, or menu item, choose Edit Link, and select the correct page or paste the verified URL. For a moved page, open the Wix dashboard, go to SEO Dashboard > Tools and settings > URL Redirect Manager, select Add Redirect, enter the old path and new path, then save.
Check whether Wix's caching or publishing state has delayed the visible change. Test the public URL in a private browser window, not only through the editor preview. If the link is inside a reusable header or footer element, update that shared component rather than editing individual pages and leaving duplicate broken references elsewhere.
Regardless of the platform, keep a redirect map with the old URL, new URL, reason, date, and owner. That record makes future migrations safer and helps identify whether a new failure comes from a content decision, a template, or a platform setting.
Testing Repairs and Maintenance Best Practices
A repair isn't complete when the CMS says “saved.” Re-crawl the source page and destination, confirm the final URL returns 200 OK, and inspect the full redirect path. If the old URL redirects, it should reach the relevant canonical page directly rather than passing through several intermediate addresses.
Schedule recurring crawls and review Google Search Console for newly reported errors. Add link checks after migrations, permalink changes, product removals, template updates, and major content pruning. A lightweight redirect register and an owner for each change are more useful than an unassigned spreadsheet full of unresolved URLs.
Ongoing governance matters because a 2026 synthesis of link-rot research reports permanent link rot rising from 5% in 2012 to 15% by 2025, with broken reference links appearing on 23% of news pages and 54% of Wikipedia pages. Treat those figures as a reason to maintain a process, not as a justification for indiscriminate redirects.
When managing crawl controls alongside link repairs, keep your robots.txt rules documented and test them with a robots.txt generation and checking workflow. A blocked destination and a broken destination are different problems, and confusing them can make diagnosis harder.
Use AutoSEO as one option for coordinating technical SEO tasks, including prioritized audit issues, redirects, internal linking checks, publishing integrations, and recurring measurement. Visit AutoSEO to review your site's technical workflow, organize broken-link repairs, and keep future URL changes from becoming another silent source of errors.
