Redirects cover old links only while the old domain and its settings survive. After a move, search the site for the old address, replace it with a WordPress-aware tool, and check each redirect takes one step.

The move that looked finished

A site moves to a new domain, or from HTTP to HTTPS. Redirects go in, the home page loads at the new address, and everybody gets on with their work.

Inside the pages, hundreds of links still point at the old address: images inserted years ago, menu items typed by hand, buttons in the page builder, the form's thank-you page.

They still work, because the redirects catch them. Each one costs an extra round trip, and each one depends on the old address and its redirects staying in place.

When either goes, they all break at once.

Where old addresses hide

The fourth is where a quick search misses most. Themes and page builders keep their settings packed inside the database, so the old address never appears in the ordinary page editor at all.

What an old link costs on HTTPS

After a move to HTTPS, an image still requested from an http address turns the page into what browsers call mixed content.

Chrome removes the padlock from such a page, and its information panel says the connection is not fully secure.

Scripts and stylesheets requested over plain HTTP are blocked outright, which can break a slider, a menu or the whole layout.

Since July, Chrome has labelled every plain HTTP page Not secure. A page that moved to HTTPS but still loads old images has paid for the move and lost its padlock anyway.

What an old link costs after a new domain

A hard-coded link to the old domain often passes through two or three redirects: old domain to its secure version, then on to the new domain.

Each is a round trip before anything loads, and a phone on a weak signal feels every one.

The redirects themselves live in the old domain's settings, usually on a hosting account. Cancel that account, or let the domain lapse, and every hard-coded link breaks together.

Ads deserve their own check. Google Ads can disapprove an ad whose address redirects to a different domain, so update the addresses in each ad directly.

Finding them

Search the database, not just the pages. Whoever maintains the site can search every table for the old address and see exactly where it appears.

Then crawl the live site with a link checker, a program that follows every link and lists those that redirect. It catches old addresses in menus, widgets and stylesheets as visitors actually meet them.

On an HTTPS site, the browser's developer console lists each mixed content warning on a page, with the address that caused it.

Outside the site, check the Google My Business listing, the social profiles, email signatures and directory listings for the old address.

Replacing them safely

Back up the database before changing anything.

A plain find-and-replace can break WordPress settings stored in a format that records the length of each piece of text.

A broken setting can empty a page layout or wipe a theme's options. Tools built for WordPress account for that.

Run a dry run first. A good tool reports how many changes it would make, and where, before it makes any.

Replace the full old address, protocol included, and check the variants: with and without www, and addresses written without http at the start. Then clear every cache so visitors see the corrected pages.

A worked example

A standby generator installer rebranded last year, moving from a long domain built around the founder's name to a shorter one, and switched to HTTPS at the same time.

Redirects were set up on the old domain's hosting account, and everything appeared to work.

This summer the owner cancelled that hosting account, which seemed to be doing nothing.

Overnight, photographs vanished from forty older project pages, the menu link to maintenance plans led to an error, and the quote form's thank-you page went nowhere, so Analytics stopped counting quote requests.

The redirects were rebuilt the next day. Then a database search found the old domain in more than a thousand places, a WordPress-aware replace changed them, and a crawl confirmed no internal link still redirected.

The form setting and the brochure PDF were corrected by hand. The old domain still redirects, now as a safety net for links elsewhere rather than the thing holding the site together.

Checking the redirects themselves

Every old address should reach its exact new equivalent in a single step: old page to matching new page, not to the home page, and not through a chain.

Test all four versions of the old address, with and without www, on HTTP and HTTPS. The secure versions need a valid certificate, or visitors see a warning before any redirect happens.

Use permanent redirects, which tell search engines the move is for good. For a new domain, add the new address to Search Console and use its change of address tool.

The counter-case

A small site with a dozen pages and no page builder can be fixed in an hour with a crawl and the ordinary editor, without any database tool.

A site due to be rebuilt within months can wait, since the rebuild will produce clean links, provided the redirects stay alive meanwhile.

And links on other sites will always depend on redirects, which is why the old domain stays registered and redirecting after the site's own links are fixed.

For everything you control, though, replace the old address rather than relying on redirects to hide it.

After any move

  1. Back up the database before changing anything.
  2. Search the database for the old address.
  3. Replace it with a WordPress-aware tool.
  4. Crawl the site for links that still redirect.
  5. Fix forms, PDFs and stylesheets by hand.
  6. Test that each redirect takes one step.
  7. Keep the old domain registered and redirecting.

Step four is the check that proves the work is done. A database search shows what was replaced; a crawl shows what a visitor still meets.

Another change best tried on a copy of the site before the live one is covered in the Try Gutenberg button.


Frequently asked questions

Why fix old links if redirects work?

Each redirect adds a round trip before the page loads, and all of them depend on the old domain and its settings surviving. Lose either and they break together.

What is mixed content?

An HTTPS page that loads images, scripts or stylesheets from plain HTTP addresses. Chrome removes the padlock for insecure images and blocks insecure scripts and stylesheets outright.

Where do old addresses hide in WordPress?

In page content, menus, widgets, theme and page builder settings, stylesheets, form settings and documents such as PDFs, many of them invisible in the ordinary editor.

Is a database find-and-replace safe?

Not a plain one. Some WordPress settings record the length of each piece of text, and a naive replace breaks them. Use a WordPress-aware tool, after a backup and a dry run.

How should the redirects be set up?

Permanent, single-step, page to matching page, covering the old address with and without www on both HTTP and HTTPS, with a valid certificate on the old domain.

Can the old domain go once links are fixed?

No. Links on other sites, old listings and printed material still depend on its redirects. Keep it registered and redirecting long after the site's own links are fixed.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Moved domains or to HTTPS this year?

Send us the address. We will crawl the site and list every link still pointing at the old one.

Start a Conversation