Every site audit I’ve run has turned up the same quiet problem. Links pointing at pages that no longer exist.
Here’s the part most posts skip. Broken internal links aren’t the same as the 404s Google tells you to ignore. Google’s guidance covers dead URLs, but these are links you placed yourself, and Googlebot keeps following them.
I’ve been doing technical SEO for 12 years, and the mistake I see most hasn’t changed. Teams chase the 404 count in Search Console while one broken footer link repeats on every page of the site.
So do they hurt you? Not with a penalty. They waste crawl paths and strand pages, and the fix is mostly process. Crawl the site, fix template-level breaks first, update links at the source instead of redirecting, and recheck every month.
Below, I cover whether they actually hurt SEO, where they come from, which tools catch them, and exactly how I run Screaming Frog to find and fix them.
What is a broken internal link?
A broken internal link is a hyperlink from one page on your website to another page or resource on the same domain that no longer loads. Instead of the destination, users and crawlers hit an error, most often a 404 Not Found, though it can also be a 410 Gone, a 5xx server error, or a redirect loop.
The “internal” part is what makes it your problem. You can’t stop another site from deleting a page, but every link on your own domain is yours to fix. That includes the ones people forget: image files, PDF downloads, and links hardcoded into the nav and footer.
Two close cousins are worth knowing. A link pointing to a 301 isn’t broken, but it adds a hop every time it’s crawled. A soft 404 is a dead page that still returns a 200 status, so crawlers can miss it unless you set them up to look. For quick definitions of terms like these, there’s the SEO glossary.
Do broken internal links hurt SEO?
Not directly. Google’s own help page on 404 errors says “404s don’t harm your site’s indexing or ranking,” and I believe them. A dead URL on its own won’t get you penalised.
The damage is indirect, and it stacks up in three places.
Link equity stops at the dead end. Every internal link passes some value to the page it points at. When that page is gone, the value goes nowhere, and the page you actually wanted to rank never gets it.
Pages quietly go orphaned. If a broken link was the only path to a page, crawlers lose their way to it. I’ve seen this after a category gets renamed: the posts under it still exist, but nothing on the site points to them anymore.
Money pages leak leads. A visitor who clicks “See pricing” and lands on a 404 doesn’t hunt for the right URL. They leave.
You’ll also read that broken links waste crawl budget. On a 200-page site, that’s mostly noise. Google only starts rationing crawls on very large sites, so don’t sell it to a client as the headline risk.
There’s a newer reason to care, though. AI crawlers like GPTBot and PerplexityBot follow your links to work out what you cover. Every dead end is a page they never read, and a page they never read is one they can’t cite.
So the 404 itself is harmless. The broken path to it isn’t.
What causes them?
Almost all of them trace back to one habit: changing a URL without adding a redirect. The page moves, the old links stay put, and nobody notices until a crawl flags it.
Here’s where I see it happen most:
- Slug edits after publishing. Someone tidies /best-crm-tools-2025/ into /best-crm-tools/ for the new year. Every link to the old version breaks overnight.
- Deleted or merged posts. Content pruning is good practice. Pruning without updating the posts that linked to the deleted page is how you end up with dozens of dead ends.
- Migrations and rebuilds. New theme, new CMS, new URL structure. This is the biggest source by far, because hundreds of URLs change in one go.
- Hardcoded nav, footer and sidebar links. One bad link in a template repeats on every page. Your crawl shows 400 errors, but it’s a single fix.
- Environment leftovers. Links to a staging domain, old http:// URLs, or a missing trailing slash your server doesn’t redirect.
- Images and files. Replace a PDF or rename an image in the media library and the old file URL dies with it. More on handling image files in my image SEO guide.
Notice the pattern? None of these are technical failures. They’re everyday edits made by someone who didn’t know a link pointed there.
Which tool should you use to find them?
Pair a crawler with Search Console and you’ll catch nearly all of them. No single tool sees everything, because each one looks from a different side. A crawler sees the links that exist on your pages. Search Console sees the dead URLs Google ran into. Analytics sees the ones real visitors hit.
| Tool | What it sees | Cost | Best for | Blind spot |
|---|---|---|---|---|
| Screaming Frog | Every crawlable link on your pages | Free up to 500 URLs, paid licence above | Deep audits | Pages nothing links to |
| Google Search Console | 404s Googlebot hit | Free | Confirming what Google sees | Lags behind, and doesn’t always show the linking page |
| Ahrefs / Semrush Site Audit | Scheduled cloud crawls | Paid subscription | Ongoing monitoring with alerts | Crawl limits set by your plan |
| GA4 404 tracking | Dead URLs visitors landed on | Free | Breaks that cost real traffic | Needs a custom event on your 404 template |
| WordPress link checker plugins | Links scanned from your dashboard | Mostly free | Small WordPress sites | Background scans add server load |
In Search Console, the list lives under Indexing > Pages > “Not found (404).” Run a URL Inspection on any entry to see where Google found the link.
I’d skip the WordPress plugins on any site where speed matters. A crawler running on your own machine does the same job without touching your server.
If you only pick one, use a crawler. Here’s exactly how I run mine.
How I find broken internal links with Screaming Frog
This is the workflow I run on every audit, whether the site has 300 URLs or 30,000. Configure first, filter to internal 4xx errors, export the inlinks, then sort by where on the page each link sits. Template-level breaks show up in minutes.
Before you crawl: set up the configuration
Five minutes here saves you from a noisy crawl. These are the settings I change before hitting Start:
- User agent: Googlebot (Smartphone). Configuration > User-Agent. Google crawls mobile-first, so I want the links Google actually sees.
- Images and PDFs ticked. Configuration > Spider > Crawl. A broken image is still a broken link.
- Parameter URLs excluded. Configuration > Exclude. A regex like
.*\?.*keeps the crawl out of filters and tracking URLs. - Speed lowered on shared hosting. Configuration > Speed. I drop to 2 or 3 threads so the crawl doesn’t slow the site for real visitors.
- JavaScript rendering, only when needed. Configuration > Spider > Rendering. Switch it on if your links load through JS. It needs the paid licence and crawls slower.
- Link Positions checked (optional). Configuration > Custom > Link Positions. Screaming Frog tags links as navigation, header, footer or content by default. If your theme uses unusual class names, adjust the rules here so step 5 works.
[Screenshot: Configuration menu with Spider > Crawl settings open. File: screaming-frog-crawl-configuration.webp]
Step 1: Run the crawl
Enter the root domain and hit Start. The free version stops at 500 URLs, which covers most small business sites.
[Screenshot: URL bar and progress bar mid-crawl. File: screaming-frog-start-crawl.webp]
Step 2: Filter to internal 4xx errors
Open the Response Codes tab and set the filter to Internal Client Error (4xx). This is your list of dead destinations, not yet the links pointing at them.
[Screenshot: Response Codes tab filtered to Internal Client Error (4xx). File: screaming-frog-internal-4xx-filter.webp]
Step 3: See where each broken link lives
Click any URL, then open the Inlinks tab in the bottom pane. You get the source page, anchor text and link position for every link pointing there.
[Screenshot: Inlinks pane with From, Anchor and Link Position columns highlighted. File: screaming-frog-broken-link-inlinks.webp]
Step 4: Export everything in one go
Clicking through URLs one at a time doesn’t scale. Bulk Export > Response Codes > Internal > Client Error (4xx) Inlinks gives you every source and destination pair in a single file.
[Screenshot: Bulk Export menu path expanded. File: screaming-frog-bulk-export-4xx-inlinks.webp]
Step 5: Spot template-level breaks
This is the step most tutorials skip, and it’s where the real time savings sit. Open the export in Google Sheets and build a pivot table: destination URL in rows, link position in columns, count of source pages as values.
A destination with hundreds of inlinks from the footer or navigation is one template fix, not hundreds of edits. [Your example: “On a recent audit, X of Y broken links came from a single footer link.”]
[Screenshot: Google Sheets pivot table of the export. File: broken-links-pivot-table-link-position.webp]
Step 6: Check redirects and server errors
Repeat the export for Internal Redirection (3xx) Inlinks and Server Error (5xx), then open Reports > Redirects > Redirect Chains. Links pointing at redirects aren’t broken, but updating them to the final URL removes a hop on every crawl.
[Screenshot: Redirect Chains report. File: screaming-frog-redirect-chains-report.webp]
Step 7: Confirm the fixes
Once the fixes are live, switch to Mode > List, paste the source pages you changed, and crawl them again. I don’t close the ticket until those pages show zero 4xx inlinks.
[Screenshot: List mode crawl with clean results. File: screaming-frog-list-mode-recrawl.webp]
One honest limit: Screaming Frog only finds links it can reach. Pages nothing links to won’t show up, so cross-check the crawl against your XML sitemap or Search Console.
Which ones should you fix first, and how?
Start with the templates, then work down by inlink count. And wherever you can, fix the link itself instead of adding a redirect.
This is the triage order I use:
- Template-level breaks in the nav, header, footer or sidebar. One fix clears hundreds of errors.
- Broken links on money pages, like pricing, services, contact and your top landing pages.
- Everything else, sorted by how many pages link to each dead URL.
Then match each link to a fix:
| Situation | What to do |
|---|---|
| Page moved to a new URL | Update the link to the new URL |
| Old URL has backlinks or traffic | Update the link and add a 301 from the old URL |
| Page deleted, nothing replaces it | Remove or rewrite the link, and let the URL return a 404 |
| Link points to a redirect | Update it to the final destination |
Why update instead of redirect? A 301 treats the symptom. Updating the link fixes the cause, and it’s one less redirect to untangle the next time the site changes. I treat 301s as a safety net for backlinks and bookmarks, not as the fix for my own links.
One trap to avoid: bulk-redirecting every dead URL to the homepage. Google’s own guidance flags this as a soft 404, so you swap a clear signal for a confusing one.
On WordPress, Yoast Premium’s redirect manager or the free Redirection plugin handles 301s without touching .htaccess. For updating links in bulk, a search-and-replace plugin like Better Search Replace saves hours. Back up the database first.
How do you stop them coming back?
Add a redirect every time a URL changes, and crawl on a schedule. Links will keep breaking, so the goal is catching them in days, not months.
- Redirect at the moment of change. Yoast Premium can create the redirect automatically when you edit a slug. If your setup can’t, make it a line on your publishing checklist.
- Crawl monthly, and after every release. Screaming Frog’s paid licence can schedule crawls, so nobody has to remember.
- Keep a redirect map. A simple sheet with old URL, new URL, date and reason. It’s the first thing you’ll want during the next migration.
- Crawl staging before a migration goes live. Catching broken links before launch is far cheaper than cleaning them up after.
- Make sure your 404 page returns a real 404 status, not a 200, and links back to your main sections.
None of this needs a tool upgrade. It needs someone who owns it. I cover technical SEO habits like these in more depth in my SEO course.
Conclusion
The issue was never the 404 itself. It’s the trail of links still pointing at pages that moved or disappeared, usually after a routine edit nobody flagged.
Preparing the fix takes one afternoon. Run a configured crawl, export the 4xx inlinks, and pivot them by link position so you know which breaks are template-level and which are one-offs.
Then fix in order: templates first, money pages next, the long tail last. Update links at the source, save 301s for URLs with backlinks, and put a monthly crawl on the calendar.
Broken links aren’t a tooling problem. They’re what happens when nobody owns the URLs. So who owns yours?
FAQ
Do broken links on my own website hurt SEO?
Broken links on your own site don’t trigger a Google penalty, because Google says 404s don’t harm indexing or ranking. They still cost you indirectly. Link equity stops at the dead page, pages can become orphaned, and visitors who hit a 404 on a pricing or service page usually leave.
I changed a blog post URL and now old links are broken. What should I do?
Update every internal link that points to the old URL so it points to the new one. Then add a 301 redirect from the old URL if it has backlinks or still gets traffic. A Screaming Frog crawl, or a search of your CMS for the old slug, shows every page that needs editing.
What’s the best free tool to find broken links on a website?
Screaming Frog SEO Spider is the best free option for most small sites. The free version crawls up to 500 URLs and shows the source page and anchor text for each broken link. Pair it with the Google Search Console Pages report, which lists the 404 URLs Googlebot found.
How do I find broken links in WordPress without a plugin?
Crawl the site with Screaming Frog from your own computer instead of installing a link checker plugin. Filter Response Codes to Internal Client Error (4xx), then export the inlinks. This skips the server load that background-scanning plugins add, which matters if you care about Core Web Vitals.
Should I redirect broken links to the homepage?
No. Google treats mass redirects to the homepage as soft 404s, which sends a confusing signal. Redirect a dead URL to the closest relevant page if one exists. Otherwise, let it return a 404 and remove the internal links pointing to it.
How often should you check a website for broken links?
Check monthly on most sites, and run an extra crawl after every migration, redesign or bulk content cleanup. Sites that publish or edit daily benefit from weekly scheduled crawls in Screaming Frog’s paid version, Ahrefs or Semrush.
Is a link pointing to a 301 redirect a broken link?
No, a link pointing to a 301 redirect is not broken, but it’s worth updating. Each redirect adds an extra hop for users and crawlers, and redirect chains slow pages down and can fail when one step changes. Point internal links straight at the final URL.