Broken Links: How to Find Every One on Your Site, and What to Do About Each

Somewhere on your website there is a link that no longer works. Probably several. A price list you replaced, a partner whose site closed, a product page renamed in a redesign two years ago, a "Download our brochure" button pointing at a file nobody moved across.

You will never find them by clicking, because the ones that break are the ones nobody clicks, right up until a customer does.

This guide covers what broken links actually cost (less than most SEO articles claim, and more than nothing), how Google really treats a 404, the redirect chains that quietly slow your pages, and how to scan a whole site for free in one sitting.

Four Different Problems, All Called "Broken Links"

They need different fixes, so it helps to separate them:

• An internal link to a page that no longer exists. Entirely your problem, entirely in your control, and the most damaging to a visitor who is halfway to buying.
• An external link to a site that has changed or closed. Not your fault, still your page. Kenyan sites accumulate these quickly, because small suppliers' websites come and go.
• A link that works but goes somewhere wrong. A staging address, an http:// address on a secure page, a # that does nothing, a javascript: link that fails on a phone.
• A link that works but takes four hops to get there. Nothing looks broken. The page is just slower than it needs to be.

What Google Actually Says About 404s

This is where most advice overstates the case, so here is Google's own position.

Google has stated that URLs which return 404 do not, by that fact alone, hurt a site or count against it in search results, and that the thing worth investigating is a page you care about returning 404. (Google Search Central, "Do 404 errors hurt my site?", dated 2011.) The same point came up again in Google's December 2023 SEO office hours: the thing to look out for is a page you consider important suddenly returning 404.

So no, a handful of 404s in your Search Console report is not a penalty, and you do not need to chase every one. What you do need to know is which pages are 404ing, because that is the difference between a dead link somebody else wrote and your own best-selling product page having quietly disappeared.

The Error That Matters More: the Soft 404

A soft 404 is a page that tells the visitor it does not exist while telling Google the opposite. The screen says "Sorry, page not found". The response code says 200, meaning success. Google's crawling documentation defines it exactly that way and treats it as something to fix. (Google: Troubleshoot crawling errors)

Google's prescribed fix depends on the page:

• Gone, with nothing replacing it: return 404 or 410.
• Moved: a 301 redirect to the new address.
• Fine, but it failed to render for Googlebot: check it with URL Inspection in Search Console.

Soft 404s are common on website builders and on WordPress sites where a plugin sends every missing address to a tidy "page not found" page that answers 200. Worse is the redesign habit of redirecting every old address to the homepage: visitors arrive nowhere near what they clicked, and Google can treat it as a soft 404 anyway.

What Broken Links Really Cost You

Set ranking aside and the costs are concrete:

Our own site is not exempt: an outside link checker once flagged broken and unsafe new-tab links on wpfoss.ke, and we fixed them in a single commit. If it happens to a site run by people who build sites for a living, it is happening to yours.

While scanning for dead links, three related problems show up on almost every Kenyan site we test, and all three are worth fixing in the same sitting:

• Links to a staging or development address. A build goes live, and three links still point at dev.yoursite.co.ke. Visitors hit a login wall or a half-finished copy of your site.
• Insecure links and mixed content. An http:// image or script on an https:// page, which browsers block or flag.
• Links that open a new tab without rel="noopener". The page you link to gets a handle on yours, which is a real, if small, security problem, and trivially fixed.

Redirect Chains: the Slow Kind of Broken

A redirect is a server telling the browser "not here, try this address instead". One is normal. Four in a row means four round trips before a single pixel appears, which on a Kenyan mobile connection is a visible delay.

Google follows up to 10 redirects in a row, which is also the limit our own checker uses. But the limit is not the target: from about three hops a chain is worth collapsing.

The chains we see most often, in this order:

1. http:// to https://, then non-www to www, then the old address to the new one. Three hops that could be one.
2. A 302 where a 301 belongs. 301 and 308 say "permanent", which is what you mean when a page has actually moved. 302, 303 and 307 say "temporary", so the old address keeps its weight.
3. A loop. Address A redirects to B, which redirects back to A. The browser gives up and nobody sees the page.
4. www and non-www ending in different places. Two versions of your site, each with its own history, and search engines unsure which is real.
5. http that never reaches https. The plain address loads, so the padlock never appears and nothing enforces encryption.

Our free redirect checker follows one address hop by hop, shows each status code, destination and how long it took, and then says in plain words what is wrong: loops, long chains, temporary where permanent belongs, and whether your www and non-www addresses agree. One honest limit worth knowing: it follows only server redirects (HTTP 3xx). A redirect done by JavaScript or an HTML meta refresh tag is invisible to any such check, including ours.

How to Scan the Whole Site at Once

Checking by hand does not scale past about five pages. Our free broken link checker reads your home page and your sitemap, then crawls your internal pages, up to 500 pages and 2,000 different links, and tests each link to see whether it actually answers. Nothing to install, no signup, and it changes nothing on your site.

Two design decisions in it are worth explaining, because they change how you should read the results:

• It says "could not verify" rather than "broken" when it cannot judge a response. Plenty of sites block automated requests or answer oddly while working perfectly in a browser. Telling you to delete a working link would be worse than telling you to go and look.
• It understands Cloudflare's email obfuscation. Cloudflare rewrites mailto: links into an encoded address that answers 404 when requested directly. A naive checker reports every email address on your site as broken. Ours decodes it and treats it as the email link it is.

It also reports the quieter problems in the same pass: links with no text at all (invisible to a screen reader and to Google), "click here" and "read more" as link text, images with no alt text, a # or empty address, internal links marked nofollow, and the staging, insecure and new-tab issues above.

What to Do With Each One

The Monthly Habit

Links rot on their own, so this is maintenance, not a project. A sensible rhythm for a small business site:

Frequently Asked Questions

Do broken links hurt my Google ranking?

Google has said that URLs returning 404 do not by themselves hurt a site in search results, and repeated the point in December 2023 office hours: what matters is a page you consider important suddenly returning 404. The damage from broken links is mostly to visitors, who leave, and to pages that should have been redirected when they moved.

What is a soft 404?

A page that tells the visitor it does not exist while telling Google everything is fine, because it returns a 200 success code. Google treats this as an error to fix: return 404 or 410 if the page is gone, or a 301 redirect if it moved. Many website builders and WordPress themes create soft 404s by sending missing pages to a styled not-found page that answers 200.

Should I redirect a deleted page or let it 404?

Redirect it if there is a page that genuinely replaces it, for example an old product moved to a new address. Let it 404 if nothing replaces it. Redirecting everything to the homepage is worse than a clean 404, because the visitor arrives nowhere near what they clicked and Google may treat it as a soft 404.

How many redirects in a row are too many?

Google follows up to 10 redirects in a row, but each hop is another round trip before anything appears on screen, so from about three hops a chain is worth collapsing. The usual cause is years of layered rules, for example http to https, then non-www to www, then an old address to a new one.

How can I check all the links on my website at once?

Use a crawler rather than clicking. The free WPfoss broken link checker reads your home page and sitemap, follows internal pages up to 500 pages and 2,000 different links, and reports which are broken, which could not be verified, and which have other problems such as links left pointing at a staging site. No signup, and it changes nothing on your site.