Broken Links SEO: What to Fix and What to Ignore

I have run broken link audits on a few hundred sites, and the conversation afterwards has two predictable outcomes. Either the owner wants to fix all 412 errors because 412 sounds like a lot, or the owner has been told 404s do not matter and now ignores the six that are quietly costing traffic and links.

Both positions are wrong in the same way. A count is not a priority list, and the decision in front of you is not find and fix. It is fix, redirect or leave, judged on evidence for each individual URL.

This is how I sort a crawl report in practice, what I do with each category, and the monthly routine that keeps link health from becoming an annual panic two weeks before a migration.

Not every 404 is a problem

Why a 404 is a valid answer to a dead URL

A 404 status code means one thing: the thing you asked for is not here. For a URL that no longer exists and has nothing worth preserving, that is the correct answer, and it resolves the situation cleanly. Google treats it that way. Nothing about a 404 asks for an error state in your site health, and a site with a small number of deliberate 404s is a normal, healthy site.

The reason owners panic is that crawl tools put a big red number next to the word error. That number describes the state of the site, not the quality of it.

The 404s that actually cost you something

A 404 costs you when one of three things is true: something links to it, either from your own pages or from another website, people are still trying to reach it, or it used to serve the answer to a search query your business cares about.

Those three conditions turn a dead URL from a non issue into money. Everything else on the list can stay as it is.

How to find the broken links that matter

Internal links first, external backlinks second

These are two separate problems and they need two separate fixes. Internal broken links are links on your own pages pointing at URLs that no longer resolve, and they cost you crawl efficiency and user experience, because you are sending both a crawler and a customer into a dead end.

Broken backlinks are links from other websites pointing at URLs on your site that no longer return a page, and they cost you the authority those links were passing. Per Ahrefs' guide to finding and fixing broken links, "broken links and backlinks are two entirely different problems."

Fix internal ones first, because they are entirely under your control and usually take an hour with a spreadsheet.

Sorting by links and traffic instead of by count

Export the crawl and add three columns: does anything link to this URL, has it had impressions or clicks in Search Console in the last 90 days, and is there an obvious equivalent page that now serves the same intent.

Then sort by those columns rather than by the size of the list. In most audits, somewhere between five and twenty URLs carry almost all of the value, and the other few hundred are noise that will consume your month if you let them.

Fix, redirect or leave: a decision rule

URLs with no traffic, no links and no replacement

Leave them. That is the whole instruction. A URL with no incoming links, no impressions, no clicks and nothing on the site that covers the same ground does not need a redirect, because a redirect to an unrelated page is worse for a user than a clean 404.

Write "leave" next to it in your spreadsheet, with the date. That single habit stops the same URLs being re-debated every quarter.

URLs earning links or visits right now

These get worked on this week. If the URL has external links pointing at it, the goal is to keep that value inside your site rather than let it land on a dead end. Redirect it to the closest equivalent page, or rebuild it if the content is worth having.

If the URL has impressions but no page, you have a demand signal and a dead end at the same time, which is the most expensive combination on the list.

Redirects done properly

Permanent and temporary, and what Search treats as canonical

Visitors cannot tell the difference between redirect types, but search systems do. Per Google Search Central's documentation on redirects, "While your users generally won't be able to tell the difference between the different types of redirects, Google Search uses certain types of redirects as a signal that the redirect target should be canonical."

In practice that means use a permanent redirect for a URL that has moved for good, and a temporary one only where the move really is temporary. If you are unsure, treat the change as permanent, because that matches how a changed URL almost always behaves six months later.

Chains, loops and internal links pointing at redirects

Three faults cause most redirect trouble. Chains, where A redirects to B which redirects to C, which slows everything down and weakens the signal. Loops, where two URLs redirect to each other, which leaves the crawler and the user nowhere. And internal links that still point at the old URL, which is the easiest of the three to fix and the one people forget.

After any redirect round, re-crawl and look for chains and for internal links to redirected URLs. A tidy redirect is a single hop from an old URL to a live one, with your own navigation already pointing at the live one.

Soft 404s: the error hiding behind a success status

What Google classes as a soft 404

This is the category that catches people out, because the page looks fine to a visitor and returns a success code to a crawler. Per Google Search Central's guidance on crawling errors, "A soft 404 error is when a URL that returns a page telling the user that the page does not exist and also a 200 (success) status code."

It matters because the status code is read first. Per Google's documentation on HTTP status codes, "For Google Search, an HTTP 2xx (success) status code doesn't guarantee indexing." A success code keeps the URL eligible and lets a useless page sit in the index, while a proper 404 or 410 closes the question.

Typical causes on WordPress and JavaScript sites

The common cause on any platform is an empty template that renders with a 200. Search pages, filtered category pages, out of stock products, and anything that returns "no results" without changing the status code. Per the same Google reference, "If the content suggests an error for Google Search, an empty page or an error message, Search Console will show a soft 404 error."

On WordPress this is usually a theme or plugin returning the archive template with nothing in it. On JavaScript driven sites it is often a client side route that renders an empty shell after the HTML has already been sent with a success status. Both are fixable, and both need the status code corrected rather than the page hidden.

Broken backlinks and the value you can reclaim

Finding pages other sites link to that you no longer serve

This is the most commercially interesting part of the work and the part most businesses skip. Sort your backlink profile by the pages receiving links, then look for URLs with real links that no longer return content.

An old service page, a discontinued product, an event page, a guide you deleted in a redesign. Each of those is a link somebody earned on your behalf that is now pointing at nothing.

When rebuilding the page beats redirecting it

If the linked page served a distinct purpose that your site no longer covers, rebuild it. A redirect hands the link to a page about something else, which is not what the linking site was pointing at, and you lose the chance to rank for the topic again.

If the topic is genuinely covered by a better page now, redirect straight to it. The rule of thumb I use: redirect when the intent is the same, rebuild when the intent is different.

What to leave alone

Internal search results and parameter URLs

Internal search result pages, filtered URLs and parameter variations are usually not worth redirecting. They were generated by a visitor's query, they should not be indexed, and they are usually best handled by keeping them out of the index rather than by writing redirect rules for thousands of combinations.

The exception is a parameter URL that has picked up external links or a steady stream of impressions. Those are worth a look individually, which is the same test you applied everywhere else.

Old URLs with nothing pointing at them and no equivalent

Old tag archives, abandoned author pages, discontinued product variants and draft URLs that were never published. Leave them as clean 404s and move on. Every hour spent redirecting a URL that nobody links to and nobody visits is an hour not spent on the five that matter.

A monthly link health routine that takes 20 minutes

• Open the crawl report and the Search Console error lists side by side
• Filter for URLs with links or impressions in the last 90 days
• Write a decision next to each one: fix, redirect or leave
• Check the redirects you created last month for chains and loops
• Search your own content for links pointing at redirected URLs
• Look at your backlink profile for important pages that no longer resolve
• Record the decisions with the date so nothing gets re-debated

Twenty minutes a month, plus a proper pass straight after any migration, redesign or platform change, is enough for most sites. The migration pass is not optional, because that is when hundreds of URLs change at once and when the expensive mistakes happen.

If the list has grown large enough that you cannot tell which errors matter, an SEO audit that separates the errors that matter is what you want rather than a longer spreadsheet. The the technical checks behind crawl and status errors cover the rest of the groundwork, and if you would rather have the sorting done for you, getting a crawl reviewed by an SEO consultant starts with the export you already have. You can also send us your crawl report for a read and get the decisions written next to each URL.

Frequently asked questions

I have hundreds of 404s. Do I need to fix all of them?

No. Google treats a 404 as a valid response for a URL that no longer exists, and the crawl report is a list of what exists, not a list of faults. Sort by evidence instead: URLs with links pointing at them, URLs with recent traffic or impressions, and URLs that sit in a navigation path or sitemap. Fix those. For the rest, a small number of deliberate 404s is a normal, healthy site.

What is the difference between broken links and broken backlinks?

Broken links are links on your own pages that point to dead URLs, and they cost you crawl efficiency and user experience. Broken backlinks are links from other websites pointing at URLs on your site that no longer return a page, and they cost you the authority those links were passing. Ahrefs makes the same split, noting that broken links and backlinks are two entirely different problems, which is why they get different tools and different fixes.

Should I redirect every old URL to the homepage?

No, and it is one of the most common mistakes in a migration. A blanket redirect to the homepage tells Google nothing about the missing page, tends to turn a specific search demand into a generic one, and leaves users without the content they asked for. Redirect to the nearest equivalent page where one exists, and let the URL return a genuine 404 where it does not.

What is a soft 404 and why does it matter?

A soft 404 is a page that tells the user the content does not exist while still returning a 200 success status, which is exactly the pattern Google describes in its crawling documentation. It matters because the status code is what the crawler reads first: a 200 keeps the URL eligible for indexing and lets a useless page sit in the index, while a proper 404 or 410 resolves it. Search Console surfaces these under its soft 404 report, and the fix is almost always to return the correct status.

My site returns an empty page when a product is out of stock. Is that a problem?

It can be, because an empty page or an error message returned with a success status is the pattern Google says will be reported as a soft 404. That does not mean out of stock pages must be deleted. It means they should return a page with real content: what the item is, when it may return, and links to alternatives. Only pages that will never carry content again should return a 404 or 410.

How often should I check for broken links?

Monthly is enough for most sites, plus immediately after any migration, redesign or platform change. A short routine works better than an annual project: check the crawl and Search Console error reports, action anything with links or traffic, clean up redirect chains created by earlier fixes, and record the decisions so the same URLs do not get re-debated next quarter.