Redirect Best Practices: A Checklist You Can Actually Follow
Most redirect problems come from a handful of repeated mistakes. Here is the checklist, ordered by how much damage each item prevents.
On this page
Almost every redirect problem traces back to one of eight mistakes. Here they are as a checklist, ordered by how much damage each one prevents, so you can work down it and stop when the remaining items stop applying to you.
1. Redirect to the Equivalent Page, Not the Homepage
The most expensive mistake, and the most common during migrations.
Mapping every dead URL to your homepage clears the 404 report, which feels productive. What it actually does is send a visitor who wanted a specific article to a page that does not contain it, and hand every inbound link a destination unrelated to what it was pointing at. Search engines frequently treat this as a soft 404 and index nothing.
Map each URL to its closest live equivalent. Where none exists, the honest options are a relevant category page or an actual 404, which at least tells the truth. Google’s site move guidance is built entirely around this mapping.
2. Match the Status Code to the Intent
301 for permanent. 302 for temporary. The mistake is using them by habit rather than by meaning.
A 301 tells search engines to transfer ranking signals and tells browsers to cache the redirect hard. That last part is what bites: repoint a 301 later and returning visitors keep landing on the old destination, sometimes for weeks, and you cannot clear their cache for them.
So anything you might change later, campaign links especially, should be a 302. Anything genuinely permanent should be a 301, because a permanent move left on 302 delays the ranking transfer indefinitely.
3. Keep It to One Hop
Every hop is a full round trip before the visitor sees anything, and a place a rule can be wrong.
The chain that catches everyone is protocol and hostname handled by separate rules: HTTP to HTTP www to HTTPS www to HTTPS. Four requests, all individually correct.
Fix it by pointing the first rule at the final destination rather than at the next rule in the sequence. Check the result with the redirect checker, which shows every hop rather than just where you ended up.
One hop is the target. Two is fine. Three should go on the list.
4. One System Owns Each Kind of Redirect (and Loops)
Two rules pointing at each other produces a browser error page and no useful diagnostic. The usual cause is two systems disagreeing: a plugin canonicalising to www while a server rule canonicalises away from it. Each is doing exactly what it was told, which is why neither looks wrong when you inspect it alone.
The prevention is ownership. Decide where redirects live and stick to it. Plugin, htaccess, Nginx config, CDN rules: any of them work, and the trouble starts when two of them are both managing the same kind of redirect.
Server level is faster and survives plugin failures. A plugin is easier to manage, searchable, and gives you click counts per rule. Pick per category rather than mixing arbitrarily: for instance, canonical host rules at server level, content and campaign redirects in the plugin.
5. Purge Caches and Test From a Clean Browser
The most frequent “my redirect is not working” report is a caching problem, not a redirect problem.
Page caches store redirect responses. So do CDNs. So do browsers, especially for 301s. Change a rule and the old response can keep being served from three different places at once.
DevDome Redirect Manager auto-purges ten common cache plugins when a rule changes, which removes the layer people forget most often. Purge your CDN too.
Then verify from somewhere clean, because your normal browser is the worst place to check a redirect. It may hold a cached 301 from before your change, it is logged in when your visitors are not, and it carries extensions your visitors do not have. A private window, or a different device on mobile data, tells you what a real visitor actually gets. If the old behaviour persists in your normal window but not a private one, that is your own browser holding the memory rather than a failed fix.
6. Write Down Why Each Rule Exists
The least exciting item and the one that pays off longest.
Two years from now, a list of two hundred rules with no notes is a list nobody will touch, because deleting an unexplained rule feels risky and keeping it costs nothing visible. So the list grows forever and eventually contains contradictions.
A one line note per rule saying what it is for and when it was added makes pruning possible. Combined with per-rule click counts, you can confidently retire anything that has served zero requests in a year and been documented as belonging to a campaign that ended.
The Quarterly Pass
Run the checklist properly once, then keep it light. The first pass is the long one, because it is where you find the chains and the contradictions that accumulated while nobody was looking. After that the list stays short, because most of what goes wrong with redirects goes wrong at the moment a rule is written rather than slowly over time.
Every quarter: check your top entry pages for accidental chains, check your campaign and affiliate links still land where intended, and look at the click counts for rules serving nothing. That is twenty minutes and it catches the drift before it becomes a project.
Disclosure: DevDome publishes this blog and makes the products it mentions. We describe what they do rather than promise results.
Key takeaways
- Redirect to the closest equivalent page, never reflexively to the homepage.
- 301 for permanent moves, 302 for anything you plan to change later.
- Collapse chains: one hop is the target, three is a problem to fix.
- Purge caches after every rule change or your fix will appear not to work.
- Document why each rule exists, or nobody will dare delete it later.
Sources
- Google Search Central: redirects and Google Search — how Google interprets redirect types and chains
- Google Search Central: site moves with URL changes — the mapping approach recommended for migrations
Links last checked August 28, 2026.
Frequently asked questions
How long should I keep an old redirect?
Permanent moves deserve permanent redirects. There is no expiry date at which an old URL stops receiving traffic, because links on other people's sites and in old emails never get updated. The cost of keeping a rule is close to zero, and the cost of removing one is a 404 on a page that still has inbound links.
Is it bad to have hundreds of redirects?
Not inherently. Large sites run thousands without trouble. What causes problems is unmanaged volume: overlapping rules, chains built from rules that each point at another rule, and a list nobody understands well enough to prune. Volume with documentation is fine; volume without it is a liability.
Should redirects live in the plugin or in the server config?
Server level is faster because the request never reaches WordPress, and it survives plugin problems. A plugin is easier to manage, gives you click counts and a searchable list, and does not require server access. The important rule is to pick one place per kind of redirect so two systems never disagree.
What is the most common redirect mistake?
Redirecting everything to the homepage during a migration. It removes the 404s from your report, which feels like success, and it strands every visitor and every inbound link at a page that does not contain what they wanted. Search engines often treat it as a soft 404 anyway.