Understand what is broken
A broken page is a URL that does not provide the content someone expected. It may return a 404 Not Found response, a 410 Gone response, a server error, or a redirect to an unhelpful destination. The first two responses can be perfectly appropriate: a site needs a way to say that content no longer exists. The more useful question is whether visitors and search engines are still being sent to that URL unnecessarily.
A soft 404 is different. The server may return a 200 OK response, suggesting a normal page loaded, while the page itself says the content is missing or shows little beyond an error message. An expired listing that displays “not found” but still returns 200 is a common example. A page that redirects every missing URL to the homepage may create a similar mismatch between what the visitor requested and what they receive.
Build a list of URLs to investigate
Start with a crawl of your own site and collect internal links that lead to error responses. Check navigation, related-content links, image links, and links generated by templates. If you use a search engine’s site management tools, review reported not-found and soft-404 examples there too. Server logs, when available, can help reveal missing URLs that people or crawlers request but that your site crawl cannot reach.
Do not treat every reported URL as a defect. A mistyped address or an old external link may point to a page that never existed. Prioritize URLs linked from important pages, included in your XML sitemap, receiving meaningful visits, or intended to be available. Keep a simple record of the requested URL, where you found it, its response code, and what the visitor sees. That makes patterns easier to spot and fixes easier to verify.
Check the response and the page together
Open each priority URL as a visitor would, then check its HTTP status code with a browser’s developer tools or a URL-checking tool. Look at the full path of any redirects, not just the final page. A URL can appear to load successfully after being sent somewhere unrelated, while the original problem remains.
For suspected soft 404s, compare the response code with the visible content. Does a missing article display an error message while returning 200? Does a product page show only an empty template? Also check whether the same template affects many URLs. A sparse page is not automatically a soft 404: a short page can still answer a specific question. Assess whether the page genuinely serves its purpose before deciding to remove or rewrite it.
Choose a fix that matches the intent
If the page should exist, restore it or correct the publishing, routing, or server issue that prevents it from loading. If the page moved, use a permanent redirect to the closest relevant replacement and update internal links to point directly to the new URL. If a link contains a typo, fix the link itself rather than relying only on a redirect.
If content is permanently gone and there is no close replacement, return a genuine 404 or 410 response. Keep the error page helpful: explain plainly that the requested content is unavailable and offer navigation or search options. Do not redirect every missing URL to the homepage. That can confuse visitors who were looking for a specific answer.
When many URLs are affected, fix the source of the problem. For example, a faulty template may keep generating links to deleted items, or a sitemap may still list removed pages. Updating the template or sitemap is more durable than correcting individual URLs one at a time. Review canonical tags on pages that should exist as well; they should not point to missing or unrelated destinations.
Verify the repair and keep watch
After making changes, test a sample of affected URLs again. Confirm that restored pages return a successful response and show the intended content, redirects reach a relevant page, and removed pages return the expected 404 or 410. Crawl the site again to check that internal links and sitemaps no longer send visitors to broken destinations.
Make this a routine maintenance task, especially after migrations, redesigns, or large content updates. Reports from search tools may take time to reflect a change, so use direct response checks to confirm the immediate result. The goal is not to eliminate every 404. It is to make sure useful pages work, missing pages are described honestly, and people can follow your site’s links without reaching avoidable dead ends.
