Use this website redirects guide to choose the right status for changed URLs, build a reliable redirect map, prevent redirect chains, and run recurring broken-link audits after a migration, redesign, or content cleanup.
Overview
A URL change is rarely just a routing task. It can affect bookmarks, campaign links, internal navigation, analytics, crawlers, and the search visibility built around the original address. The safest approach is to decide what the old URL means now, map it to the most relevant destination, and verify the result before and after launch.
Redirects send a browser or crawler from one URL to another. A canonical URL is different: it is a signal that identifies a preferred version among pages that remain accessible. A soft 404 is different again: it is a page that appears successful at the server level but effectively tells visitors that the requested content is unavailable. Treating these mechanisms as interchangeable creates confusing behavior and makes audits harder.
Redirect reference table
| Mechanism | Typical use | Key caution |
|---|---|---|
| 301 | A permanent move, such as changing a slug, consolidating content, or moving to a new domain. | Use only when the old location should consistently resolve to a new one. |
| 302 | A temporary change, test, short-term campaign route, or temporarily unavailable destination. | Do not use it as a default for a permanent migration. |
| 307 | A temporary redirect that preserves the original request method and body. | Confirm that the application and destination handle the request as intended. |
| 308 | A permanent redirect that preserves the original request method and body. | Use with care when redirecting non-GET requests, such as form submissions or API calls. |
| Canonical URL | Indicating a preferred URL when multiple accessible URLs represent similar or duplicate content. | It is not a replacement for a redirect when the old URL should no longer be used. |
| Soft 404 | Not a redirect type; it describes an unavailable page that returns a normal-looking success response or a generic empty page. | Return an appropriate not-found response or redirect only when there is a genuinely relevant replacement. |
For most permanent page moves, a 301 is the familiar starting point. For temporary routing, evaluate whether a 302 or method-preserving 307 is more appropriate. A 308 can be useful in systems where preserving the request method is required, but test it carefully before applying it broadly.
What to track
1. Create a redirect map before changing URLs
A redirect map turns an informal migration into a reviewable deployment artifact. Record the old URL, its proposed destination, redirect type, reason, and test result. Include URLs from the current sitemap, analytics or server logs, important backlinks, campaign documents, and internal navigation.
Use a structure such as this:
| Old URL | New URL | Status | Reason | Owner | Test |
|---|---|---|---|---|---|
| /old-guide | /guides/new-guide | 301 | Slug changed | Web team | Pending |
| /summer-offer | /offers/current | 302 | Temporary campaign route | Marketing | Pending |
Do not automatically send every retired URL to the homepage. A redirect is most useful when the destination matches the visitor's likely intent. If no relevant replacement exists, a clear not-found page is often more useful than a misleading redirect.
2. Track redirect quality, not only redirect counts
For each rule, check the response status, final destination, protocol, hostname, path, and whether the page loads correctly. A redirect that technically resolves may still be wrong if it lands on an unrelated page, removes a required path segment, creates a loop, or changes a query parameter used by the application.
Monitor these recurring indicators:
- Redirect coverage: important old URLs have a deliberate destination or a documented not-found outcome.
- Final status: the destination returns the expected response instead of another redirect or an error.
- Chain length: the old URL reaches its final destination with as few hops as practical.
- Loop detection: no rule sends a URL back to itself or into a repeating sequence.
- Internal links: menus, articles, XML sitemaps, canonical tags, and structured references point directly to current URLs.
- Broken-link volume: new 4xx responses are investigated rather than allowed to accumulate.
3. Keep related infrastructure in scope
URL changes often occur alongside domain, hosting, or HTTPS changes. Confirm that DNS points to the intended environment, that both HTTP and HTTPS behavior are deliberate, and that the certificate covers the hostnames visitors use. For related checks, see the Domain DNS Setup Guide and the SSL Certificate Renewal Guide.
Redirect rules should also be stored with the project configuration where possible. This makes them reviewable, repeatable, and easier to restore if a deployment is rolled back. On a managed cloud host, check whether rules belong in the application, web server, hosting panel, or edge configuration before editing them.
Cadence and checkpoints
Run the most detailed audit around a launch, then use a lighter recurring review. The following cadence works for many small sites and can be adjusted to match publishing and deployment frequency.
Before a migration or redesign
- Export the current URL inventory from the CMS, sitemap, analytics records, logs, and important campaign materials.
- Classify each URL as keep, move, merge, remove, or investigate.
- Assign a destination only where the replacement is relevant and available.
- Review the map for conflicting rules, wildcard patterns, query-string behavior, and trailing-slash differences.
- Save a backup and establish a rollback plan before deploying changes.
At launch
Test representative URLs from every rule group, including an old page, a nested path, a URL with parameters, a page that should remain unchanged, and a URL that should return not found. Test from the public hostname rather than only from a local development environment. Confirm that redirects work over the intended protocol and do not expose a temporary or staging host.
Immediately crawl the new site or sample it systematically. Review server responses, final URLs, page titles, canonical references, navigation, forms, and key conversion paths. Update internal links and the sitemap to use final URLs directly; relying on redirects inside your own site adds unnecessary requests and makes future audits less precise.
Monthly or quarterly review
On a monthly or quarterly cadence, review new 3xx and 4xx responses, top landing URLs, recently deleted content, and redirect rules added since the previous check. A lightweight developer utility, log report, or scheduled crawl can make this repeatable. Keep a dated issue log containing the URL, first-seen date, response sequence, likely cause, action, and verification date.
How to interpret changes
A rise in redirects is not automatically a problem. It may reflect a planned migration or a new content structure. The useful question is whether the changes are intentional, efficient, and relevant to visitors.
Fix redirect chains
A chain occurs when URL A redirects to URL B, which redirects to URL C. Chains commonly appear when a site has undergone several naming or protocol changes. Replace the sequence with a direct rule from A to C when A is still receiving traffic or may be linked externally. Also update internal links so they target C directly.
Watch for mixed chains such as HTTP to HTTPS, old domain to new domain, and old slug to new slug. These may be valid individually but inefficient together. Test the complete sequence from the original public URL and record every response in the audit log.
Separate routing errors from content decisions
If many old URLs redirect to one generic destination, inspect the map rather than assuming the rule is successful. A broad rule may hide deleted content, misspellings, or a broken CMS route. Conversely, a large number of 404 responses after a deliberate cleanup may be acceptable if those URLs have no relevant replacement and the not-found page helps visitors continue.
Canonical tags deserve a separate review. If two accessible pages are substantially similar, confirm that the preferred URL is consistent in navigation, sitemap entries, redirects, and canonical markup. Do not add a canonical tag merely to conceal a broken route or substitute it for a required permanent redirect.
Use tools carefully
Command-line requests, browser developer tools, hosting logs, CMS reports, and site crawlers can all reveal different parts of the redirect path. Compare results when behavior differs by user agent, cookie, authentication state, or request method. If redirect rules use regular expressions, test edge cases before deployment; the Regex Testing Guide can help document reusable test cases.
When to revisit
Return to this guide before every URL-structure change, domain move, CMS migration, hosting migration, and major content consolidation. Revisit the redirect map after publishing campaigns expire, products or services are retired, or a site section is renamed. A quarterly broken-link audit is a sensible baseline for a small site; review more often when URLs change frequently or deployments are continuous.
Update the process when your web server, hosting panel, CMS routing model, browser behavior, or search documentation changes. The implementation examples and validation workflow should be checked at least every 6–12 months, even if the site's URL structure has remained stable.
A practical maintenance checklist
- Export new 3xx and 4xx findings.
- Compare them with the previous audit and the approved redirect map.
- Remove obsolete temporary rules and document any permanent exceptions.
- Replace redirecting internal links with direct links.
- Test representative old URLs, current URLs, parameters, HTTP/HTTPS variants, and not-found cases.
- Check that destination pages load correctly and preserve the visitor's intended task.
- Record the fix, test date, and person or system responsible for verification.
For broader launch preparation, pair this process with a Website Backup and Restore Testing Guide, the Website Uptime Monitoring Guide, and the shared-hosting-to-managed-cloud migration checklist. A redirect audit is most valuable when it is part of a repeatable deployment and maintenance routine, not a one-time cleanup after links have already failed.