Broken link building means finding dead outbound links on relevant pages, then offering the site owner a working replacement — usually yours. It still works in 2026, but it’s a numbers game with a low hit rate. Expect to check dozens of prospects for each link you actually earn.
Broken link building has been declared dead roughly once a year since 2013. It isn’t. What’s changed is that the lazy version stopped working: mass-scraping 404s, blasting a template at 500 addresses, and hoping. That approach now produces almost nothing except spam complaints.
The version that still works is slower and narrower. You find a genuinely useful resource that died, confirm people still link to it, build or identify something at least as good, and tell the handful of people whose pages are currently broken. That’s a real favour wrapped around a real ask.
What is broken link building?
It’s an outreach tactic where you locate outbound links on third-party pages that point to dead URLs, then contact the page owner suggesting they swap the dead link for a live resource of yours. The pitch works because you’re reporting a real defect on their page before asking for anything.
The mechanics are simple. A blogger wrote a post in 2019 citing a tool, a study, or a guide. That resource has since gone offline — the company folded, the site restructured, the author let the domain lapse. The blogger has no idea. Their page now sends readers to a 404, which is bad for them and invisible to them. You noticed. You have something that covers the same ground. That’s the entire play.
Does broken link building still work in 2026?
Yes, but the margin has narrowed. Site owners get more outreach than they did five years ago and recognise the pattern instantly, so the tactic now lives or dies on whether the dead resource was genuinely valuable and whether your replacement is genuinely better.
Two things have made it harder. First, saturation — a page with 40 referring domains pointing at a dead URL has probably received a dozen replacement pitches already, and whoever got there first won. Second, the rise of content that exists purely to be a broken-link replacement, which editors can smell from a distance. A thin 800-word article published last Tuesday is not a credible substitute for a five-year-old industry resource. What still helps: link rot is relentless, so every year produces a fresh crop of dead resources, and most people have moved on to easier tactics.
How do you actually run a broken link building campaign?
The workflow is five stages: find pages with dead outbound links in your topic, qualify whether the dead resource is worth replacing, secure a replacement that’s honestly better, find a named human to contact, and send a short specific email. Most of the value comes from the qualifying stage, which is the one everyone skips.
-
Build a prospect list
Three sources work reliably. Competitor broken backlinks: most backlink tools have a “broken backlinks” report — run it on three or four competitors and you get every page linking to a dead URL on their domain. Dead domains in your niche: find a resource that shut down (industry news, a Twitter thread, a tool directory with a dead entry), then pull the full backlink profile of that dead domain. That’s the richest source by a distance, because one dead resource can have hundreds of live pages still pointing at it. Resource page crawls: search
"your topic" + "resources"or+ "useful links", then run the resulting pages through a link checker. -
Verify the link is actually broken
Don’t trust the tool. Load the URL yourself — a surprising share of reported 404s are temporary outages or geo-blocks. Then check the Wayback Machine: if the page never had real content, the prospect is worthless.
-
Qualify the dead resource
This is the stage that decides your hit rate. Ask three things. Was it genuinely useful — a real tool, real data, a real guide — or was it a thin blog post nobody would miss? How many live pages still link to it? A dead URL with 3 referring domains is a morning wasted; one with 60 is worth building a resource for. And is the linking context specific, meaning the page describes what the resource did, or is it one entry in a 200-item link dump? Specific context converts; link dumps almost never do. Worth applying the same standards you’d apply to any placement — the way we assess whether a host page is worth having a link on works just as well here.
-
Get the replacement right
Read the archived version and be blunt with yourself: is your page a fair swap? If the dead resource was a state-by-state regulation table and yours is a general overview article, the answer is no, and the editor will see that in four seconds. Either you already have something comparable, or you build it — and if the dead URL has 60 referring domains, building it is a rational use of a week.
-
Find a named contact
Generic inboxes convert poorly. Look for the post author’s byline first, then the site’s about or team page, then LinkedIn. If the post is bylined “Sarah Chen, Editor” you want Sarah, not info@. If you can only find a contact form, use it — form submissions from a real person with a specific observation get read more often than people assume.
-
Send, then follow up exactly once
One email, one follow-up seven to ten days later, then stop. A second follow-up buys you almost nothing and costs you goodwill. Track everything in a sheet: prospect URL, dead URL, referring domain count, contact, date sent, outcome. Without that record you’ll re-pitch the same person in three months.
What should the outreach email actually say?
Short, specific, and free of flattery. Name the exact page, quote the exact dead link, say what it used to be, offer the replacement in one line, and make it genuinely easy to say no. Under 120 words is a good ceiling.
A template that doesn’t read like a template
Subject: broken link on your [topic] page
Hi [First name],
I was reading your piece on [exact article title] and the link to [name of dead resource] under the “[section heading]” heading is returning a 404 — looks like [domain] went offline sometime in [year, from the Wayback Machine].
Wanted to flag it either way.
If you’re replacing it, we put together [your resource title] which covers [the specific thing the dead page did]: [URL]. If it’s not the right fit, no problem at all — the dead link is the part worth knowing about.
[Your name]
[Your site]
Change three things per prospect: the article title, the section heading where the link sits, and the phrase describing what the dead resource did. Those details prove you opened the page. If you can’t fill in the “specific thing the dead page did” slot, you haven’t qualified the prospect — skip it.
Most templates fail for one of four reasons. They open with invented praise (“I’ve been a huge fan of your blog for years”) that no editor believes. They bury the broken link under two paragraphs of preamble. They ask for the link with more urgency than they report the problem, which flips the favour into a demand. Or they’re so obviously mail-merged — no section heading, no resource name, “your recent article” — that the recipient stops reading at line two.
What conversion rate should you expect?
Low, and anyone telling you otherwise is selling something. In our experience across outreach generally, reply rates land somewhere around 1 in 5 to 1 in 8 when the email is specific and genuinely useful — and a reply is not a link. Placements from those replies are a fraction again.
Here’s the honest arithmetic on a well-qualified campaign. Say you find a dead resource with 80 referring domains. Maybe 45 of those pages are still live, relevant, and worth pitching after you strip out the scraped junk and abandoned blogs. You find contacts for 30. You get 5 or 6 replies. Two or three of those become links, and one of them turns out to be nofollow. That’s a genuinely decent result, and it took you the better part of a week including building the replacement resource.
Which puts broken link building in the same effort bracket as most manual link acquisition — commonly 4 to 6 hours of work per link that actually goes live, once you count the misses. That’s not a criticism of the tactic. It’s just what earned links cost in time, and it’s worth knowing before you build a plan around it. If that maths pushes you toward having someone else handle the prospecting and correspondence, that’s a reasonable conclusion to reach — it’s the same work either way, and the question is only whose hours it consumes.
Why do most broken link campaigns fail?
Almost always because of prospect quality rather than email quality. People spend hours optimizing subject lines while pitching pages where the dead link never mattered, the site is abandoned, or the replacement is transparently inferior.
| What went wrong | Why it kills the campaign | The fix |
|---|---|---|
| The dead resource had 4 referring domains | Even a perfect hit rate gives you one link for a day’s work | Set a floor before you start — we wouldn’t build a replacement for anything under about 25 live referring domains |
| Your replacement is thinner than what died | Editors compare against the archived version, and a 700-word overview replacing a data tool is an obvious downgrade | Read the Wayback capture first and only pitch when you can defend the swap out loud |
| Pitching link-dump resource pages | Nobody maintains a 300-link directory, and the link carries little value anyway | Prioritise pages where the dead link sits inside a paragraph that explains it |
| Emailing info@ or a generic form | Nobody owns that inbox, so nobody acts | Spend the extra three minutes finding the author’s name; skip the prospect if you can’t |
| Template too obviously mail-merged | Recipient pattern-matches to spam within one line | Include the section heading and the dead resource’s actual name — details a bot wouldn’t have |
| Site is abandoned | No one has published since 2021 and no one will edit anything | Check the most recent post date before adding to the list |
| Someone got there first | The link was already replaced weeks ago and your pitch is stale | Re-verify the 404 the same day you send, not the day you built the list |
| Three follow-ups | Turns a helpful note into harassment and burns the domain relationship | One follow-up, then move on and log it |
When is broken link building the wrong tactic?
Skip it when your niche has too little published history to have produced dead resources, when you have nothing to offer as a replacement and no capacity to build one, or when you need results inside a month. It’s a slow tactic with a long lead time between effort and outcome.
It’s also poorly suited to commercial pages. Nobody’s broken link was a pricing page, so your replacement is almost always informational — a guide, a tool, a dataset. The links land on content, and the internal linking side of the job is where a lot of hard-won external links quietly go to waste.
And there’s a scale question. If you need 15 links a quarter on a predictable schedule, broken link building alone won’t deliver that reliably — the supply of good prospects in any one niche is finite and lumpy. Most sites end up running it alongside other approaches. The trade-offs between the main options are covered in our comparison of placement types, and broken link building sits closest to a link insertion (also called a niche edit) in that it puts your link inside an existing, already-indexed article.
What tools do you need?
A backlink tool with a broken-backlinks report, a bulk link checker, an email finder, and a spreadsheet. That’s genuinely the whole stack, and the spreadsheet is doing more work than any of the software.
- Backlink index — for broken backlink reports on competitors and dead domains. Ahrefs, Semrush and Majestic all do this; the free tiers are too limited for real prospecting.
- The Wayback Machine — non-negotiable. You cannot qualify a dead resource you haven’t read.
- Email finder — Hunter, Findymail or similar, plus manual checking of about pages. Verify before sending; bounces damage your sending reputation.
- A tracking sheet — prospect URL, dead URL, referring domains, contact name, send date, follow-up date, outcome. Update it the day you send.
Key takeaways
- Broken link building works because you report a real problem before making an ask — that’s why it out-replies cold guest post pitches.
- The dead-domain approach beats scattered 404 hunting: one shut-down resource can leave hundreds of live pages pointing nowhere.
- Qualifying decides your hit rate. Check whether the dead resource was genuinely useful, how many live pages still link to it, and whether the linking context is specific.
- Always read the archived version before pitching a replacement. If your page is thinner than what died, don’t send.
- Keep the email under 120 words, name the article and section heading, and make refusing easy.
- Reply rates commonly sit around 1 in 5 to 1 in 8 on specific outreach, and replies aren’t links — budget roughly 4 to 6 hours per link that goes live.
- One follow-up, seven to ten days later. Never a second.
How many prospects do I need for one link?
Plan on contacting 8 to 15 qualified sites per acceptance, which is the same ratio that holds across most manual outreach. That’s 8 to 15 after filtering — the raw list you start from will be three or four times bigger once you strip out abandoned sites, scraped pages and prospects you can’t find a contact for.
Should I build a new page just to replace a dead resource?
Only when the numbers justify it. If the dead URL has 60-plus live referring domains and the resource was substantial, spending a week building a real replacement is rational. If it has 5, you’re building a page for one possible link, and your time is better spent elsewhere.
Do site owners actually care about broken links?
Editors at maintained sites usually do — a 404 in their content is a small, fixable embarrassment. Abandoned blogs don’t care because nobody’s reading the inbox. This is exactly why checking the last publish date before you add a prospect saves more time than any email optimization.