Blog / Technical SEO /

Diagnosing redirect chains on a site with many editors

Somebody fixes an old URL by pointing it at a new page. A year later someone else moves that page and adds another redirect. Three editors, three sensible fixes, and now a visitor takes four hops to reach a product page.

Redirect chains usually start with good intentions. With many editors the rules pile up until nobody can trace them. The visitor may still arrive, but each extra hop adds complexity and can hide a loop or a pointless destination. Diagnose the chain before you delete rules that protect old links.

Find the path

Crawl known old URLs and inspect each response, or use logs and redirect records if you have them. Record the source, destination, status and number of hops. Check whether WordPress, a redirect plugin, the server config or the CDN adds a step. A plugin’s list won’t show rules applied elsewhere. That’s the trap I watch for first.

Confirm the right destination

A shorter path isn’t better if it sends people to the wrong page. Compare the final destination with the original intent and the current content. Update internal links to point straight at the preferred URL, then plan which middle rules can be removed safely. Look for loops, and for mixed HTTP and HTTPS or hostname changes.

Document and retest

Keep a reason and an owner for each rule that stays. Test representative URLs after edits, and again during migrations. I’d rather have a short list of documented redirects than a long one nobody dares to touch.

A good rule of thumb is that every redirect should be explainable in a sentence. “The product was renamed in 2024” is fine. “Nobody knows” is a rule to investigate before you delete it.

Remember that editors rarely create chains out of carelessness. They fix what is in front of them. A shared log, even a simple one, gives the next editor the context they need to point a new rule straight at the final page.

Make the review reusable

Write down the decision the review needs to support before you start deleting rules. “Clean up the redirects” is too broad to finish.

Take one real old URL and follow it hop by hop to the final page. Record the status codes, who added each rule and why. Ask an editor for a typical case and a strange one. A method that only works for a clean two-step chain will fail on the messy ones.

If the evidence is missing or contradictory, say so and ask the person who knows the history. Don’t let a plugin or a script tidy it away.

Keep a note of what each rule did before you changed it, so a later reviewer can tell a deliberate rule from an accident. Make exceptions visible and scoped. Check nearby sections that share the same patterns, since a fix in one folder can leave the same chain in another.

Name one person who can approve a destination and another who can confirm the rule is live. Redirect work often sits in a shared queue because everybody can touch it and nobody owns it.

When you report on a chain, separate what you observed, what might explain it, and what you propose. “Three hops, added by two plugins and the CDN” is far more useful than “redirects are messy”.

Be honest about trade-offs. Collapsing every chain protects speed and tidiness, but it can erase the record of why a rule existed. Say which you’re choosing.

Don’t judge the cleanup by whether the ticket closes. Check that real URLs now land where visitors expect.

Test the rule on a representative URL and inspect the live response before applying it to a whole section.

If the evidence is limited, choose a small validation step and say what it can’t tell you. A crawl of fifty old URLs shows a pattern, but it isn’t the whole site.

Choose a review trigger tied to something that creates chains: a migration, a URL restructure, a hostname change or a new editor joining. Those are better prompts than a calendar reminder.

Ask the editors for a normal case and a meaningful exception. If your method only copes with a tidy two-step chain, it will fail on the weird ones.

Keep a record of why each rule exists and what it replaced. Without it, the next reviewer can’t tell a deliberate rule from an accident or an old export.

Revisit the chains when a migration, template or hostname changes. Which redirect chain would be risky to shorten without checking its original purpose?