A site owner opens Google Search Console and sees 300 URLs under Not found (404). The tempting response is to generate 300 redirects and make the report go away.
That is usually the wrong starting point.
A 404 simply tells the browser or crawler that a URL was not found. It does not prove a site has a ranking problem, and it certainly does not mean that every missing URL needs to point to the homepage. Google's own 404 guidance says that many missing URLs are normal. A deleted page with no replacement can keep returning 404; a page that moved to an equivalent address should usually redirect.
The real question is what happened to the page, and what should someone visiting the old address find now?
Use this decision table before creating a redirect
- The content moved to an equivalent page: redirect the old URL directly to that specific replacement. A visitor should still find what they came for.
- The original page should exist but is accidentally missing: restore it or repair the publishing route. A redirect might hide the underlying problem.
- The page was deliberately removed with no close replacement: a genuine 404, or 410 where appropriate, is a valid outcome.
- A sitemap or internal link contains a typo: fix the link at its source. Consider redirecting the typo if people still use it.
- The URL already returns HTTP 200: inspect its current content and crawl history before making any change. The reported 404 may be old.
- A page returns 200 but displays “not found”: investigate a possible soft 404. An HTTP success code alone doesn't mean the content exists.
These are decisions about individual URLs. The “all broken pages to homepage” rule is not a shortcut; it can leave visitors on an unrelated page and create a soft 404 problem. Google explains this distinction in its crawling-error documentation.
First separate historical reports from current problems
Search Console's Page indexing → Not found (404) examples are not a live test performed the moment you open the report. They describe what Google found when it crawled those URLs.
A page might have been restored yesterday. A redirect might already be working. Or the URL may still be missing, but only because someone linked to a typo years ago.
Open the URL directly or check its HTTP response before deciding what to change. If you have a long list, ArsKhan's Redirect Converter and live 404 checker can check batches from the exported Search Console examples.
Work through a Search Console export
- In Google Search Console, open Page indexing, then select Not found (404).
- Export the Examples table of URLs as CSV. The chart and Performance exports are different files; they do not give this importer the input it expects.
- Open the Redirect Converter and choose Import GSC 404s.
- Upload the CSV and run Check next 25 URLs. The checker contacts the selected public URLs and reports their current HTTP outcomes. Network rate limits may require several batches or sessions.
- Download the checked-results CSV. Separate pages that still return a direct 404 from those returning 200, redirects, errors or other statuses.
- For each direct 404, decide whether to restore it, correct a source link, redirect it to a close replacement, or leave it missing.
- Only where a relevant replacement exists, enter that destination and prepare a draft redirect rule. Verify the destination yourself before publishing the rules on your host.
The importer accepts a bounded CSV/TXT list rather than accessing your Google account. It does not change your Search Console property, publish redirects, or decide replacement pages for you. The import is processed locally; the selected URLs sent for a live check are requested by ArsKhan's server. Do not treat the result as an independent statement of what Google has indexed.
Three example URLs, three different decisions
Imagine a small technical blog that changed its structure last month. The URLs below are fictional.
1. The article moved
Old: example.com/blog/email-authentication-basics
New: example.com/guides/email-authentication-basics
The content is substantially the same and the old URL is genuinely obsolete. A permanent redirect to the new article makes sense. Update internal links and the sitemap to use the new address too.
2. A seasonal offer ended
Old: example.com/black-friday-2024-coupon
There is no equivalent offer today. Sending that URL to the homepage just to remove a 404 would mislead visitors. Leaving a proper 404 is reasonable, provided you remove stale internal links or sitemap entries that advertise the page.
3. A live article is accidentally missing
Old: example.com/guides/install-on-linux
The article still belongs on the site, but a route change broke it. Restore the correct page or its intended URL rather than redirecting it somewhere vaguely related. The broken route may be a publishing defect worth fixing.
What matters here is not whether the URL appears in a report. It is whether the outcome serves the person who follows the link.
A 404 in a sitemap deserves a different response from a random URL
Search Console may show URLs from places you control and from places you do not. A missing address still present in your XML sitemap is a useful signal: you are actively telling search engines that it belongs on the site.
Check the current sitemap and internal links. If the page moved, replace the old entries with the canonical new URL. If it was deleted, remove it from the sitemap. ArsKhan's XML Sitemap Inspector can help identify structural problems and questionable URL entries, but it is not a substitute for confirming which pages should exist.
Google's Page indexing report help recommends focusing especially on missing pages you link to yourself or include in a sitemap. That is a much better triage rule than “redirect everything.”
What about soft 404s and redirects that already work?
A soft 404 is different from a server returning HTTP 404. For instance, a product might no longer exist but the site serves a bare “item unavailable” message with HTTP 200. Google may consider the response an error page despite the success status.
Likewise, a URL that returns 301 → 200 is not currently a direct 404. If Search Console still lists it as one, compare the crawl information with today's response. Do not stack a second redirect on top without a reason.
Some live checks can also fail because a server blocks automated requests, throttles repeated traffic or requires authentication. Treat those cases as unverified, not as confirmed missing pages.
Decide what gets fixed first
If there are hundreds of examples, prioritise the URLs with the clearest consequences:
- Important pages that should be live and still appear in your site navigation.
- Missing URLs still referenced in your sitemap or internal content.
- Old pages with known relevant replacements and valuable inbound links.
- Commonly visited misspellings that can be corrected without creating irrelevant redirects.
Leave random nonexistent paths and abandoned pages without replacements until there is evidence they matter. The goal is not a perfectly empty 404 report. It is a website where important links lead to the right content.
After the change, prove it worked
For a redirect, check the actual old URL, not just the generated configuration. Confirm that it returns the expected permanent redirect, reaches the intended page, has no loop and doesn't take visitors through needless hops. For a restored page, confirm that it returns 200 and contains the intended content.
Then monitor Search Console over subsequent crawls. A fixed URL may remain in a historic issue report for a while, and a submitted validation request does not force immediate indexing. Google does not guarantee instant crawling or inclusion in search results.
A useful 404 workflow ends when the site makes sense to visitors—not when every row of a CSV has a redirect.
Need to work through your own list? Open the Redirect Converter's GSC 404 importer, check the live status and prepare only the redirect rules you can justify.