Broken Links: Which Ones Cost You Anything

Most broken links cost you nothing. Fixing every one of them is a waste of a developer’s afternoon. But one category — external sites linking to URLs on your domain that now return 404 — quietly destroys link equity you already paid for, in outreach, content, or PR. The job is not “find all broken links”; it is “find the broken links that transfer value, and triage the rest.”

First, Get the Vocabulary Right

SEO teams lump four different problems under “broken links,” and they behave nothing alike:

• Internal links that 404 — a page on your site links to another page on your site that no longer exists. This is a crawl and UX bug.
• Outbound links that 404 — your page links to someone else’s dead page. Mostly cosmetic.
• External backlinks pointing at your dead URLs — other sites link to a page you deleted or moved. This is the expensive one.
• Redirect chains and loops — technically not broken, but they leak crawl budget and dilute signals along the way.

A broken link checker finds all four indiscriminately. The triage — deciding what actually matters — is yours.

The Classification That Decides Where You Spend Hours

Here is the working table. Bookmark it; most broken-link audits go wrong because the team treats every row the same.

Type	Impact	Priority	How to fix
Internal links to a 404	Medium-high: wasted crawl budget, frustrated users, diluted internal PageRank flow	Fix immediately	Update the link to the correct live URL, or remove it. Redirect only if the content genuinely moved.
Outbound links to a 404	Low: minor trust/UX signal, no direct ranking penalty	Fix during routine content updates	Replace with a live equivalent source, or unlink the anchor text. A redirect to a homepage is not a fix.
Backlinks pointing to your 404s	High: you are leaking accumulated link equity on pages that used to rank	Highest — act now	301 to the closest relevant live page, or rebuild the page if the URL earned many referring domains.
Redirect chains (2+ hops)	Low-medium: crawl inefficiency, slow page loads, signal dilution across hops	Fix in batches	Point the first link directly at the final destination and collapse the chain.
Soft 404s (page returns 200 but shows “not found”)	Medium: search engines may treat it as thin or duplicate content	Fix when found	Serve a real 404/410 status, or replace with genuine content.

Internal Broken Links: Annoying, Real, and Cheap to Fix

Internal 404s will not get you penalised. What they do is worse in a boring way: they waste crawl budget on large sites, break the flow of internal link equity, and signal neglect. On an e-commerce site with tens of thousands of SKUs, expired product pages left linked from category pages are the classic offender.

What to actually do

1. Crawl your own domain with a site-wide broken link checker — most crawlers cover sites of 100,000+ URLs on a standard plan, so check vendor limits against your URL count rather than trusting feature lists.
2. Fix the source. Edit the page so the link points to a live, relevant URL. This is always better than adding a redirect, because it removes a hop entirely.
3. Only redirect when the target content genuinely moved. Do not mass-redirect every dead URL to the homepage — Google treats large volumes of irrelevant redirects as soft 404s, which means you have achieved nothing.

What not to do

Do not run this as a one-off project. Sites change. Content gets deleted, migrations go half-done, plugins rot. Set the crawl to run monthly and put internal 404s into your normal dev backlog with a threshold — for example, any crawl with more than 0.1% broken internal links triggers a ticket. A quarterly “link audit sprint” is the wrong cadence; by the time you run it, half the findings are stale.

Outbound Broken Links: Mostly a Non-Problem

Here is the popular advice that is wrong: “Broken outbound links hurt your rankings.” There is no credible evidence that a dead external link on your page costs you positions. What it costs you is reader trust — someone clicks your citation, hits a 404, and quietly discounts everything else you wrote.

Treat outbound 404s as a content-quality task, not an SEO emergency:

• Fix them during scheduled content refreshes, not as standalone work.
• Replace the dead source with the closest live equivalent — an updated study, an archived version, a newer official page.
• If no equivalent exists, remove the link and keep the text. An unlinked claim beats a citation to a graveyard.

The exception: pages you actively use for link building. If you are pitching guest posts or digital PR, dead references on those pages undermine your pitch credibility with editors who check. Prioritise those.

Backlinks to Your Own 404s: The Category That Costs Real Money

This is the row in the table that justifies the whole exercise. When Site A links to your guide at /guides/old-topic and that URL now 404s, several things happen:

• You get the brand mention but none of the ranking benefit — the equity stops at the dead URL.
• The referring site’s editor eventually notices the dead link and removes it. Now you have lost the link outright.
• If the page used to rank and earned those links over years, deleting the URL threw away its entire link profile.

The scale of the loss depends on how many referring domains point at the dead URL. A 404 with two weak backlinks is a shrug. A 404 with 40 referring domains is a page you should probably rebuild today.

The triage process

1. Pull your backlink profile from a link index and cross-reference the target URLs against live pages on your site. Your broken link checker finds the dead URLs; the link index tells you which ones carry weight.
2. Sort dead URLs by number of referring domains. Set a threshold: anything with 3+ referring domains gets individual attention; anything below gets a redirect batch.
3. For high-value dead URLs, decide: rebuild the page at the same URL (best — you keep the URL’s history), or 301 to the closest live relevant page (acceptable), or let it die (only if it’s truly obsolete and the links are low quality).
4. Relevance matters more than convenience. Redirecting /seo-audit-checklist to your homepage passes little useful signal because engines and users treat irrelevant redirects as soft 404s. Match topic to topic.

Budget realistic time: a mid-size site audit with a few thousand referring domains typically takes one person a full working day to triage properly, plus developer time for redirects. Compare that against the cost of acquiring equivalent new links through outreach and the maths is usually lopsided.

Bonus: Turn Other People’s Broken Links Into Yours

The same mechanics run in reverse. Broken link building works like this: find pages in your niche with dead outbound links, create (or point to) content that replaces the dead resource, and pitch the site owner to swap the link to you.

It sounds elegant. In practice, success rates are low, and honest practitioners say so. Expect low single-digit response rates on cold outreach — often under 5% — because:

• The site owner has no relationship with you and no reason to act.
• Many dead links point to resources that no longer have a good replacement, so you would have to build content from scratch for one prospect.
• Webmasters are inundated with exactly this pitch template, and it reads as spam.
• The volume maths is brutal: to win 10 links at a 3% conversion rate, you need roughly 330 qualified prospects, each verified manually.

It still works when the pitch is specific, the replacement is genuinely better than the dead resource, and you target pages updated recently enough that the owner still maintains them. It fails as a scaled, templated campaign. If an agency promises broken link building at volume, ask for their actual placement-to-prospect ratio.

Choosing a Broken Link Checker: Evaluate the Job, Not the Feature List

Vendors change pricing, crawl limits, and features constantly, so instead of listing products, here is what the tool must actually do — check any candidate against this:

• Full-site crawl at your scale. A checker that samples 500 pages is useless on a 50,000-URL site. Confirm crawl limits against your real URL count.
• Backlink cross-reference. The killer feature: mapping external referring domains to your dead URLs. Without this, you cannot do the high-priority triage above. Some standalone checkers do not do it at all — you need a link index for that half.
• Status code detail. You need to distinguish 404 from 301, 302, and soft 404s (200 with “not found” content), not just a binary “broken” flag.
• Scheduled monitoring. Manual one-off scans go stale within weeks. Alerts on new broken links beat quarterly panic audits.
• Exportable reports. You will hand findings to developers. If the tool cannot export a clean URL list with source pages and status codes, someone retypes it by hand.

Pricing across the market ranges from free single-page checkers to enterprise suites costing hundreds of dollars per month. What drives cost is crawl volume and link-index access, not the link check itself, which is trivially cheap. Pay for scale and the backlink mapping; do not pay for a prettier dashboard.

A Workable Cadence, Summarised

• Monthly: automated full crawl; internal 404s ticketed and fixed within one sprint.
• Quarterly: backlink-to-dead-URL cross-check; anything with 3+ referring domains triaged individually, rest redirected in batch.
• During content refreshes: outbound links verified and replaced.
• After every site migration or content deletion: immediate crawl plus redirect plan before, not after, the old URLs go live-dead. Post-migration is where most link equity is permanently lost, because nobody notices the 404s until rankings have already slid for 4–8 weeks.

The discipline is simple: fix internal links fast, ignore outbound links until you touch the page anyway, and treat every backlink pointing at your own 404 as money left on the table. Most broken links are noise. That last category is the signal.

Related reading

• AI search optimisation — scope, method and what we measure.
• Pricing
• Why Every Backlink Tool Shows a Different Number
• Website Change Monitoring: Catching Silent SEO Regressions