The Page That Redirects Perfectly Into a Wall
Every old URL 301s. The redirect map was thorough. Nothing 404s.
Six weeks later, traffic to a significant share of those pages has evaporated, and the redirects were part of the reason.
Not because they failed technically — they resolve correctly. Because of where they point. A redirect is a statement that one URL has been replaced by another, and search engines evaluate whether that claim is credible. When the destination bears no resemblance to the source, the claim isn't credible, and the redirect is treated much like a soft 404.
Relevance Determines Whether Signals Transfer
The mistake is treating a redirect as a pipe. Point the old URL anywhere, and whatever the old URL had flows through.
It doesn't work that way. Search engines compare the old page's content and context against the new destination. If they're closely related, the redirect consolidates signals. If they aren't, it's treated as a page that no longer exists.
The catch-all-to-homepage pattern is the clearest example. Mapping hundreds of deleted product pages to the homepage feels tidy — nothing 404s, everything resolves. In practice, a redirect from a specific product page to a generic homepage is an irrelevant destination, and these are commonly treated as soft 404s.
The same applies to category-level dumping. Redirecting fifty discontinued products to one category page is better than the homepage, and still weak if the category doesn't actually answer what those pages answered.
The ordering to aim for:
- Direct equivalent. The same content at a new URL. Full consolidation.
- Closest topical match. A page covering substantially the same subject.
- Relevant parent category. Weaker, but defensible where no direct match exists.
- A 410 Gone. For content genuinely removed with no successor.
- Homepage. Essentially never the right answer for a specific page.
Option 4 is underused. A 410 is an honest signal — this content is gone deliberately. It removes the page from the index faster than a 404 and doesn't waste crawl budget on a redirect to something irrelevant.
Building the Map
The Redirect Mapper helps plan and audit the mapping:
- Enter your source and destination URLs.
- Review the mapping.
- Check for chains and loops before deploying.
The process that produces a good map:
1. Inventory everything. Not just what's in your CMS. Crawl the live site, export URLs from analytics, pull URLs from search console, and check server logs for URLs receiving traffic. Old URLs with external links pointing at them won't necessarily appear in a crawl.
2. Rank by value. Pages with external backlinks and pages with organic traffic matter most. Map those first and most carefully. A long tail of zero-traffic pages can be handled with patterns.
3. Map individually where value is high, by pattern where it isn't. A regex rule handling /blog/2019/(.*) → /articles/$1 covers hundreds of URLs. Individual mapping for the twenty pages that carry most of the value.
4. Validate every destination. Each target must return 200, be indexable, and be self-canonical. A redirect to a noindexed page or to another redirect defeats the purpose.
5. Check for chains and loops before deploying, not after.
Chains Accumulate Across Migrations
A chain is a redirect pointing at another redirect:
/old-page → /newer-page → /current-page
It resolves. Each hop costs latency, and search engines may not follow long chains fully.
Chains form the same way every time: a migration's redirects are written against the URLs that exist today, and today's URLs include the destinations of the previous migration's redirects.
Three site restructures over five years, each adding a layer, produces chains of four or five hops that nobody deliberately created.
Flatten rather than append. When adding new redirects, update every existing rule that points at a URL you're now moving, so it points directly at the final destination:
Before adding:
/v1-page → /v2-page
Moving /v2-page to /v3-page. Add:
/v2-page → /v3-page
And update:
/v1-page → /v3-page ← flattened, not left as a chain
This requires treating your redirect rules as a maintained dataset rather than an append-only log. Which is exactly why a mapping document that outlives any single migration is worth keeping.
Loops are the pathological case — A redirects to B, B redirects back to A. Browsers detect this and show an error after a number of hops. Usually caused by two independently-written rules interacting, or by a trailing-slash rule fighting a canonical-hostname rule.
The Rules That Fight Each Other
Several common redirect rules interact badly when written independently:
Trailing slash normalisation. A rule adding trailing slashes and another rule stripping them creates an immediate loop.
HTTPS enforcement. http:// → https:// should happen in the same hop as any hostname change, not as a separate redirect after it.
Hostname canonicalisation. example.com → www.example.com should combine with the protocol upgrade:
Bad (3 hops):
http://example.com/page → https://example.com/page → https://www.example.com/page → https://www.example.com/page/
Good (1 hop):
http://example.com/page → https://www.example.com/page/
Case normalisation. Uppercase-to-lowercase rules combining with path rewrites.
Language detection redirects. Geographic or header-based redirects on top of everything else — and these can prevent crawlers reaching versions they're trying to index.
The fix is ordering and combination. Write one rule that produces the final canonical form in a single hop, rather than several rules each fixing one attribute.
301, 302, 307, 308
Choosing the wrong status code has real consequences.
301 Moved Permanently. The default for a permanent move. Signals to search engines that the new URL replaces the old one and that signals should be consolidated. Cached aggressively by browsers — sometimes indefinitely — which is why a mistaken 301 is painful to undo.
302 Found. Temporary. Signals should stay with the original URL. Use for genuine temporary situations — an A/B test, a seasonal page, maintenance.
307 Temporary Redirect. Like 302 but preserves the request method. A POST stays a POST. Historically, some clients converted POST to GET on a 302.
308 Permanent Redirect. Like 301 but preserves the method.
410 Gone. Not a redirect. Signals deliberate permanent removal, and removes the page from the index faster than a 404.
The common error is using 302 for permanent moves. It's often a default in frameworks and CMS plugins, and it means signals don't consolidate onto the new URL.
The dangerous error is 301 for something temporary, because browsers cache it and users may keep hitting the redirect long after you've removed the rule.
Redirects Take Time
Deploying a redirect doesn't immediately move rankings.
Search engines must recrawl the old URL to discover the redirect, then process the move, then reassess the destination. This takes time proportional to how frequently the old URL is crawled — days for high-traffic pages, potentially much longer for obscure ones.
Practical implications:
Don't panic in week two. A traffic dip immediately after a migration is expected. Judge at six to eight weeks.
Keep redirects in place indefinitely. Removing them after six months breaks every external link and every bookmark that still points at the old URL. Redirects are permanent infrastructure, not migration scaffolding.
Monitor the old URLs, not just the new ones. Search Console will show you which old URLs are still being crawled and how they're being handled.
Update internal links to the new URLs. Internal links pointing at redirects waste crawl efficiency and add latency for users. The redirect stays for external traffic; internal links should point directly.
Practical Tips
Map to relevance, not to convenience. A homepage redirect for a specific page is close to a deletion.
Use 410 where content is genuinely gone rather than redirecting somewhere loosely related.
Combine protocol, hostname and path rules into a single hop.
Flatten chains when adding new rules, don't append to them.
Validate every destination returns 200 and is self-canonical.
Keep the redirect map as a maintained document that outlives any one migration.
Update internal links after deploying redirects.
Test before deploying — a loop in production is an outage for those URLs.
FAQ
Can I redirect everything to the homepage? Technically yes, and it usually doesn't work. Irrelevant redirects are commonly treated as soft 404s and don't consolidate signals.
What's the difference between 301 and 308? Both are permanent. 308 preserves the request method; 301 historically allowed clients to change POST to GET.
When should I use 410 instead of a redirect? When content has been deliberately removed and no relevant equivalent exists. It's an honest signal and removes the page from the index faster than a 404.
How long should redirects stay in place? Indefinitely. External links and bookmarks don't expire, and removing a redirect breaks them.
Why did my traffic drop after a migration? Some short-term fluctuation is normal while search engines process the moves. Persistent drops usually mean redirects pointing at irrelevant destinations, chains, or destinations that aren't indexable.
Are redirect chains actually harmful? They add latency, may not be followed completely, and accumulate silently across migrations. Flatten them.
The Takeaway
A redirect is a claim about equivalence, and it only works when the claim holds up. Mapping to the closest genuine match — and using 410 honestly where no match exists — matters more than ensuring nothing 404s. The tidy-looking catch-all to the homepage is the version that quietly loses the most.
Plan and audit URL redirects free with the Redirect Mapper at sadiqbd.com — no sign-up, instant results.
