404s waste budget and frustrate users. The fix is usually a redirect — but not always, and doing it wrong is worse than leaving it.
Find them
- Check the Pages report in Search Console for 404s Google found
- Watch your server logs for repeated 404 hits
- Crawl your own site periodically and extract links
- Check Search Console → Links → Top linked pages for broken targets
Redirect or let it 404?
- Redirect (301) when the content moved or a close equivalent exists
- Let it 404/410 when the content is gone with no replacement — a 404 is better than a redirect to something unrelated
- 410 signals "gone permanently" and can be deindexed faster than 404 in some cases
Avoid chains and loops
Point redirects directly to the final destination. Chains (A → B → C) dilute signals and slow users down.
- Audit for chains: each redirect should have one hop
- Never redirect A → B and B → A
- Never redirect to a 404
The mechanics
- Apache:
Redirect 301 /old /new - Nginx:
return 301 /new;inside alocationblock - Most CMSs and hosts have a redirect UI
- Next.js:
redirects()innext.config.js
Nginx example
location = /old-page {
return 301 /new-page;
}Apache example
Redirect 301 /old-page /new-pageWhere redirects matter most
- Site migrations (domain, CMS, URL structure)
- Product/category renames in e-commerce
- Blog slug changes
- Removing seasonal or discontinued pages
Worked example
After moving a blog from /2023/03/seo-tips to /blog/seo-tips, you add a single 301 from the old URL to the new one, update internal links to point at the new URL, and submit the new URL for indexing. Do not create a chain through an intermediate slug.
How to verify
curl -I https://example.com/oldshould return301with the rightLocation- Follow it:
curl -ILshould end at200 - Search Console → Pages should show the old URL as "Page with redirect"
Generate the rules
Use the free redirect generator to turn a list of old → new URLs into ready 301/302 rules for nginx or Apache, and clean up internal links with the link extractor.