Broken Links and 404 Pages: How Much Do They Affect SEO?

Broken links are one of the most misdiagnosed problems in SEO. Not because they don’t matter, but because the version that matters is rarely the version people are looking at.

The 404 count in Google Search Console is where most people start, and it is almost always the wrong place. We audit sites across everything from local service businesses to national franchises and B2B manufacturers, and the number of 404s a site is carrying has never been the thing holding its rankings back. Google’s position is the same. 404s are not a quality signal, and Google’s crawling documentation lists the belief that they waste crawl budget as a myth.

The failures that genuinely cost money sit elsewhere. A page with earned links that quietly died during a site rebuild. An empty category returning a 200 status code. A blanket redirect rule sending every retired URL to the homepage, which Google reads as a soft 404. Those are all broken link problems. None of them show up as the number people are worried about.

This post separates the two. What Google actually says, what a broken link genuinely costs, which ones to fix, and which ones you can safely leave alone.

TL:DR

• 404s are not a penalty. Google has said so repeatedly. John Mueller said in January 2026 that 404s and 410s are not a negative quality signal, and that this is how the web is supposed to work. There is no threshold, and no sitewide penalty for having thousands of them.
• 404s do not waste crawl budget. Google’s crawling documentation lists this as a myth. A 404 tells Googlebot not to bother with that URL again.
• Soft 404s do waste crawl budget. A 200 status code tells Googlebot the page is worth coming back to, so it keeps returning to a page that has nothing on it.
• The real cost is lost link equity. A link pointing at a dead page passes nothing on. That is a wasted asset rather than a punishment, and it is the main reason to audit at all.
• Broken internal links are worse than broken external ones. They are entirely within your control, and they interrupt journeys your visitors were already partway through.
• 404 and 410 are handled the same way by Google, despite what most guides claim. Use whichever your website platform makes easier.
• Redirecting everything to your homepage is worse than doing nothing. Google reads it as a soft 404, and visitors read it as being ignored.
• Fix the broken links that had traffic or backlinks. Leave the rest. A 404 on a page that should not exist is the correct answer, not a failure.

What Counts as a Broken Link or 404 Page?

These two terms get used interchangeably, and that is where most of the confusion starts. They describe different things, and the difference determines how much attention each one deserves.

404 Not Found, Explained

Every time a browser or a search engine requests a page, your server answers with a status code. A 200 means the page is there. A 404 means the server looked, and it isn’t.

That is all a 404 is: an accurate answer to a request for something that does not exist. Google treats it accordingly. Content from a page returning a 404 is ignored, and if the page was indexed previously, it gets dropped from the index over time. Google also crawls that URL less and less often.

You will see 410 mentioned alongside 404 in most guides, usually with a claim that it works faster. Google’s own documentation says otherwise, and every 4xx status code apart from 429 is handled the same way. More on that later.

Broken Links and 404 Pages Are Not the Same Thing

A 404 page is a destination. A broken link is a pointer aimed at one. You can have thousands of 404s and almost no broken links, or a handful of 404s doing real damage because of what is pointing at them.

There are three kinds of broken link, and they are not equally important:

Broken internal links. A link on your own site pointing at one of your own dead pages. Entirely within your control, and the easiest to fix.

Broken inbound links. Another website linking to a page of yours that no longer exists. This is where value is actually lost, because that link was doing work for you and has stopped.

Broken outbound links. A link on your site pointing at a dead page on someone else’s. Untidy, occasionally embarrassing, rarely urgent.

Most of the confusion around broken links and SEO comes from treating these three as one problem. The right response to each is different, and so is the cost of ignoring it.

Soft 404s: The Problem Worth Worrying About

A soft 404 is a page that tells your visitor it doesn’t exist while telling Google that it does.

The page displays an error message, or nothing much at all, but the server still returns a 200 status code. The usual culprits are a broken database connection, a missing server-side include, a script that failed to load, or a search results page with no results on it.

This is the version worth acting on, for one specific reason. A 404 tells Googlebot not to come back. A 200 tells it the opposite. So Googlebot keeps returning to a page with nothing on it, and Google is explicit that soft 404s continue to be crawled and waste crawl budget.

In practice the causes are duller than that list suggests. Empty category pages. Filtered listings that return nothing. Plugins set up to show a “page not found” template without changing the status code to match. Search Console flags them in the Page Indexing report under “Soft 404”.

So Do Broken Links Affect SEO?

Yes, but not in the way the question usually implies. There is no penalty, no threshold, and no sitewide damage from the count itself. What there is, in some cases, is a loss you have chosen not to recover.

What Google Actually Says

Google’s position on this has been consistent for well over a decade, and it has been restated recently enough that it should be settled by now.

In January 2026, John Mueller was asked on Reddit why 404s do not really matter for SEO. His answer was that a 404 means the URL simply does not get indexed, which is fine, and that 404s and 410s are not a negative quality signal. He added that this is how the web is supposed to work.

Gary Illyes made a similar point on LinkedIn in 2024, opening with the observation that 404 errors are nothing to be afraid of and that you do not need to scramble to fix them, at least not most of the time. His advice was to fix the ones that will actually help users, which is a more useful instruction than any number.

Mueller has also said that a site could have a substantial share of its URLs returning 404s without that looking unusual to Google. The one exception he raised was the homepage. If that starts returning a 404, Google may reasonably wonder whether the site still exists.

Then there is crawl budget, which is where most SEO advice on this topic goes wrong. Google publishes a page of crawling myths, and one of the entries is the claim that pages serving 4xx status codes waste crawl budget. Google’s answer is that they do not. Googlebot attempted the crawl, received a status code, and got no content to process. There is nothing there to waste.

The nuance that matters: soft 404s do waste it, because a 200 response invites Googlebot back indefinitely. Which means the standard advice has it backwards. The pages people rush to fix are harmless, and the ones they never look at are the ones burning resource.

Not all broken links carry the same cost. This is the version worth keeping.

What’s broken	Real SEO impact	Priority
A 404 on a page that should not exist	None. This is the correct response	Leave it
A 404 on a page that should exist	High. A live page is missing from search	Fix immediately
A page with backlinks now returning 404	High. Earned authority is going nowhere	Redirect to closest match
Broken internal links	Moderate. Interrupts journeys, wastes your own linking	Fix in batches
Soft 404s	Moderate. Repeatedly crawled, never indexed	Fix at template level
Everything redirected to the homepage	Moderate. Google reads it as a soft 404	Undo, redirect properly
Broken outbound links	Minimal. A tidiness issue	Housekeeping

 

Nothing in that impact column describes a punishment. Every genuine cost is either a page that should be earning and isn’t, or a link that should be passing value and isn’t. Broken links do not make Google think less of your site. They make parts of your site stop working, and you pay for that in traffic rather than in rankings.

Where Broken Links Actually Cost You

Each item below is something that was working, stopped working, and is now costing you. None of them is a penalty, and that distinction changes what you go looking for.

The Short Answer, By Link Type

Lost Link Equity Is The Expensive One

When another website links to you, that link passes value. Google treats it as a signal that your page is worth something, and it contributes to how that page ranks. This is usually called link equity, and it is the closest thing in SEO to an asset you have genuinely earned.

Point that link at a page that no longer exists and the value stops moving. The link is still there on the other site. It still gets clicked. But as far as Google is concerned it now arrives at nothing, and the authority it was passing goes nowhere.

This is the single most expensive thing on the list, and it is almost always self inflicted. Sites do not lose backlinks by accident. They lose them during rebuilds, replatforms, and URL structure changes where the redirect map was incomplete or never made at all. A page that took three years to earn its links can be retired in an afternoon.

It is also entirely recoverable. A 301 redirect from the dead URL to the closest equivalent page passes most of that value on, and Google treats a 301 as a strong signal that the destination is the right one.

An internal link is a decision you made. You linked from that page to that one because you wanted people to follow it. When the destination dies, the decision is still sitting there in your content, quietly failing.

Three things happen. Visitors who were part way through a journey hit a dead end. Any value you were passing between your own pages stops passing. And Google loses a path it was using to find and understand the destination page. None of this is dramatic, but it is a slow leak in something you built deliberately, and unlike broken inbound links, nobody else has to cooperate for you to fix it.

Soft 404s Multiply

Hard 404s appear one at a time. Soft 404s tend to arrive by the thousand, because they are almost always a template problem rather than a page problem.

One misconfigured category template, one filter that returns an empty listing instead of a proper error, one plugin serving a “page not found” message without changing the status code, and you have generated soft 404s across every URL that template touches. Because they return a 200, Googlebot keeps coming back to all of them. Which is why they are worth fixing at the source rather than one by one. Find the template, fix the status code, and the whole set resolves together.

Visitors You Have Already Paid For

The commercial cost is the one that gets least attention and is easiest to quantify.

Someone reaching a broken page has already been acquired. They found you in search, clicked a link from another site, or followed a URL from an old email, a printed leaflet or a Google Business Profile listing. The work of getting their attention is done. The cost of it is already spent. Then the page fails, and most of them leave. Not to a competitor necessarily, but away from a journey they had already started.

To be clear about what is and isn’t happening here: this is not a ranking signal. Google is not watching people bounce off your 404 page and marking you down for it. The loss is straightforwardly commercial. You paid to get someone to a page that didn’t work, and they didn’t come back.

When a 404 Is the Correct Answer

Most advice on this subject assumes every 404 is a mistake waiting to be corrected. It isn’t. A page that no longer exists should return a 404, and building elaborate machinery to avoid that causes more problems than it solves.

Pages That Should Stay Dead

Some URLs deserve nothing more than an honest error. A discontinued product. An event from four years ago. A campaign landing page that did its job. A URL somebody mistyped that never existed in the first place. A page removed deliberately because it was thin, outdated or duplicated elsewhere.

For all of these, the 404 is doing exactly what it should. It tells Google the page is gone, keeps it out of the index, and reduces how often Googlebot bothers checking. Nothing is being lost, because there was nothing left to lose.

The question worth asking is not “how do I get rid of this error” but “did this page have traffic or links”. If the answer is no, you are looking at a report entry rather than a problem.

The Homepage Redirect Trap

The most common bad fix is a rule that sends every 404 to the homepage. It is easy to implement, it makes the Search Console report look tidier, and it is worse than doing nothing.

Mueller has walked through what actually happens. The URL still does not get indexed. It may well be classified as a soft 404 anyway and carry on being crawled. And it confuses visitors, who clicked a link expecting one thing and landed somewhere unrelated with no explanation of why.

You have replaced an honest error with a misleading one. Google is not fooled, the URL is no better off, and the person who clicked is now more lost than they were before. Redirects should point at content that serves the same need as the original page. If there is a genuine equivalent, redirect to it. If there isn’t, let the 404 stand.

404 or 410? It Barely Matters

This one has been repeated so often it has become received wisdom. The claim is that 410, meaning “gone”, removes a page from Google’s index faster than a 404, meaning “not found”.

Google’s documentation does not support it. All 4xx status codes apart from 429 are handled identically, and 404 and 410 sit in that group together. Mueller has said directly that sending a 410 will not change how long a URL keeps being recrawled. Some Googlers have described 410 as marginally faster at purging a URL, and marginally is the operative word.

Use whichever your platform makes easier. If your CMS returns 404 by default, that is fine. If setting a 410 for deliberately removed pages costs you nothing, that is also fine, and it is semantically more accurate. Neither choice will show up in your rankings.

Google Will Not Forget the URL

Worth setting expectations here, because it catches people out. Google does not permanently forget a URL it has discovered. A page can return a 404 for years and still get recrawled occasionally, and it can keep appearing in your Search Console report long after you have stopped caring about it.

Mueller’s view on this is more relaxed than most people’s. Asked whether repeated crawling of 404s was a problem, he said these do not cause problems and would leave them alone, adding that in a way it means Google would be happy to pick up more content from the site. A long list of 404s in Search Console is not a backlog. It is a record.

How to Find Broken Links and 404 Pages

You need two views of this, and most people only ever get one. Search Console tells you what Google has found. A crawler tells you what is on your site right now. The gaps between those two lists are usually where the interesting problems are.

Google Search Console

Start here, because it is free, it is specific to your site, and it reflects what Google has actually seen rather than what a tool thinks it should have seen. Open the Pages report under Indexing. Below the summary you will find the list of reasons pages are not indexed, including “Not found (404)” and, separately, “Soft 404”. Click either to see the affected URLs.

Two things worth knowing before you read too much into it. The report shows URLs Google has crawled, so it will include addresses that never existed on your site, mistyped links from elsewhere, and old URLs from years ago. And it is a sample rather than a complete list, so treat the count as a signal rather than an inventory.

The genuinely useful move here is not fixing the list. It is sorting it. Which of these URLs used to get traffic, and which have other sites linking to them? Everything else can wait indefinitely.

A Crawler for Internal Links

Search Console will not tell you which of your own pages contain broken links. For that you need to crawl the site.

Screaming Frog is the standard choice and the free version handles up to 500 URLs, which covers a lot of the sites this applies to. Run a crawl, filter the response codes to 4xx, then look at the Inlinks panel for each broken URL. That panel is the part that matters, because it tells you which pages contain the broken link and therefore what you actually need to edit.

Sitebulb does the same job with more hand holding. Semrush and Ahrefs both include site audit tools that will surface broken internal links on a schedule, which is more useful than a one off crawl if nobody is going to remember to run it again.

Finding Broken Inbound Links

This is the one worth the most and the one people skip, because you cannot find it by crawling your own site. Another website is linking to a page of yours that no longer exists, and nothing on your site knows about it.

You need backlink data for this, which means Ahrefs, Semrush, or Majestic. All three will show you links pointing at URLs that return an error. Ahrefs calls this Broken Backlinks, Semrush surfaces it in Backlink Analytics.

If you have no paid tools, there is a workaround. Cross reference your Search Console 404 list against pages that used to earn organic traffic, using either Search Console’s own performance data or your analytics. Any URL that once had visibility and now returns a 404 is a candidate worth checking manually.

Analytics and Server Logs

Your analytics will tell you which broken pages people are actually reaching. In GA4 this means filtering page views by the title your 404 template uses, which requires that your 404 page has a distinctive title in the first place. Plenty do not.

Server logs are the complete picture, showing every request and every status code returned including the ones no tool sampled. They are also the most work to interpret, so they are worth reaching for when something has gone wrong at scale rather than as routine maintenance.

How to Fix Broken Links and 404 Pages

Everything so far leads to one question, asked URL by URL: did this page have traffic or links? That question decides the fix, and it is the only triage you need.

The Decision, In Four Options

Restore it. The page should exist and something removed it. A deletion in error, a database problem, a URL that changed during a rebuild without anyone noticing. Put it back at the original address. This is the only option that recovers everything.

Redirect it. The page is genuinely gone but a close equivalent exists. Use a 301 to the page that serves the same need the original did. A discontinued product goes to its replacement or its category, not the shop front. A retired service page goes to the service that replaced it.

Leave it. No traffic, no links, no equivalent. The 404 stands. This is the correct answer far more often than people expect, and it costs nothing.

Mark it gone. Same as leaving it, with a 410 instead of a 404 if your platform makes that easy. Semantically tidier. Practically identical, as covered earlier.

The failure mode to avoid is redirecting things that fall into the third category. A redirect to something unrelated is not a neutral act. Google may read it as a soft 404, and the visitor gets sent somewhere they did not ask for with no explanation.

Fix the Source, Not Just the Symptom

Redirecting a dead URL solves the destination. It does not solve the link that pointed at it.

If pages on your own site link to a URL you have just redirected, edit those pages so they link to the live address directly. The redirect works, but you have left a hop in place permanently for no reason, and long redirect chains have a real cost. This is the step almost everyone skips, and it is often the highest impact one. Your crawler already told you which pages contain the broken link. Editing them takes minutes.

The same applies to your sitemap. If dead URLs are still listed in it, you are actively telling Google to go and look at them. Regenerate it.

Recovering Broken Inbound Links

For a dead page with links from other sites, the 301 does most of the work and passes most of the value on.

Where the destination is a meaningful upgrade on what was there, it is worth going a step further and asking the linking site to update the link. This works more often than people assume, particularly with suppliers, partners, industry directories and local press, who generally have no idea the link broke and no objection to fixing it.

Platform Notes

WordPress. The Redirection plugin handles this well and logs 404s as they occur, which gives you a live list of what people are actually hitting. Rank Math and Yoast Premium both include redirect managers if you would rather not add another plugin. Check that your theme’s 404 template returns a genuine 404 status code, because some do not.

Shopify. URL redirects live under Navigation in the admin. Worth auditing after any collection restructure, and worth checking what happens to a product URL when the product is deleted rather than archived.

Anything custom. The redirect belongs in the server configuration or the application layer rather than in a plugin, and the test is the same regardless of platform. Request the URL and check the status code that comes back, not the page that renders. A page that looks like an error while returning a 200 is a soft 404, whatever it says on screen.

What a Good 404 Page Should Do

A 404 page will not improve your rankings. It is worth getting right anyway, because by the time someone sees it you have already paid to bring them there.

Get the status code right first. This is the part that matters technically and the part most often broken. The page must return a genuine 404, not a 200 with an apology printed on it. If it returns a 200, you have built a soft 404 no matter how helpful the wording is. Request the URL and check the response, rather than trusting what appears on screen.

Then make it useful. The job of a 404 page is to give someone a next step, so it needs to explain plainly that the page is missing, offer a search box, link to your main sections, and keep your normal navigation and branding in place. Someone who lands on a stripped back error page with no header has no idea whether they are still on your site.

Give it a distinctive page title as well. It costs nothing and it is what lets you count these visits in GA4 later.

What to avoid is easier to list. Do not redirect to the homepage, for the reasons already covered. Do not leave your host’s default error page in place, which tells the visitor nothing and looks broken. And do not put so much effort into the joke that the page forgets to help.

Creative 404 pages get a lot of attention because they make good screenshots. They are not what this page is for. The visitor arrived somewhere unexpected and wants to get back on track, and a clear route out is worth more than a clever illustration.

The Short Answer

Broken links affect SEO, but almost never in the way the question is usually meant. There is no penalty for having 404s, no number that triggers one, and no crawl budget being consumed by pages that return an error. Google has said this repeatedly and its own documentation lists the crawl budget claim as a myth.

What broken links actually do is stop things working that used to work. A page that was earning traffic disappears. A link somebody gave you stops passing value. A journey a visitor had already started ends in a dead end. Those are real costs, and they are worth recovering, but they are losses rather than punishments and they respond to a different kind of attention.

So the useful discipline is not clearing the list. It is sorting it. Take your 404s, work out which ones had traffic or links, fix those properly, and leave the rest returning an honest error with a clear conscience. A 404 on a page that should not exist is your site behaving correctly.

The one thing worth checking regardless of how relaxed you feel about the rest: whether anything on your site is returning a 200 status code while telling visitors the page cannot be found. That is the failure hiding behind a tidy report, and it is the one nobody goes looking for.

If you would rather someone else went through it, that is the kind of thing we do. Get in touch and we will take a look at what your site is actually returning.