You find a dead link on a page that ranks for exactly the keyword you've been trying to crack for months. The page is authoritative, it's active, and the link points to a resource that disappeared years ago. That's not a problem. That's an opening.
Broken link building step by step is one of those tactics that sounds too simple to work until you actually run it. Then you realize it works because it exploits something most outreach ignores: the webmaster already has a problem you can solve. You're not asking for a favor. You're pointing at a 404 and offering a fix. That shift in framing is the whole game.
I've run this process on sites ranging from personal blogs to mid-market SaaS. My best single campaign landed 11 links from 34 emails sent over five weeks. My worst was 0 out of 22, and it took me a while to understand why. That failure taught me more than the win did.
Key Takeaways
- Broken link building works because you lead with value (a broken link report), not a request.
- The qualification step matters more than the outreach step — target pages must be relevant and still maintained.
- The main goal is to earn a contextual link from an established page by replacing a dead resource with your own.
- Response rates hover around 5–15% in my experience; anything above 20% usually means a smaller, tighter list.
- Dead pages with no replacement option are a dead end — skip them early instead of wasting outreach.
- AI has cut the prospecting time roughly in half for me, but it hasn't improved my reply rate at all.
What is the main goal of broken link building?
The goal is to get a link placed on an existing, indexed page that already ranks and already has authority, by giving the webmaster a reason to update that page. You are not creating a new page. You are repairing an old one, and you get credit for the repair.
That distinction matters because most link building asks the webmaster to do work. Guest posting asks them to review your article, format it, publish it. Directory submissions ask them to approve an entry. Broken link building asks them to swap one URL for another. It's the lowest-effort ask in the entire discipline.
Why the mechanism works
A broken link is a real defect. If a page links to a resource that now returns a 404, the page has lost value — visitors hit a dead end, and the editor's original intent is broken. Fixing it is genuinely in their interest, independent of your request.
The reason this converts better than cold pitching is that you're doing the diagnostic work for free. You've already found the problem, identified the page, and prepared a replacement. The webmaster only has to click one link and decide yes or no.
Broken backlink building vs. other link tactics
Not all link building scales the same way or costs the same effort. Here's how I'd compare the three tactics I've spent the most time on:
| Tactic | Effort per link | Typical response | Best fit |
|---|---|---|---|
| Broken link building | Medium — prospecting heavy | 5–15% reply rate in my runs | Sites with strong resource pages |
| Guest posting | High — writing required | 10–25% but slow | Building topical authority |
| Digital PR | Very high — creative + outreach | 2–5% but high authority | Brand visibility campaigns |
Broken link building sits in the middle. It's more mechanical than PR, less creative than guest posting, and it produces contextual links from pages that already have trust signals. That's the sweet spot.
So which methods are not actually backlink building?
People conflate a lot of things under the link building umbrella. Buying links, link exchanges at scale, and comment spam are the obvious ones — they're tactics, but they're the kind that get you penalized rather than ranked. Less obviously, submitting your site to a directory that auto-approves every submission isn't link building either. It's noise. The link exists, but it carries no editorial weight, which is the entire point of the exercise.
Broken link building is on the opposite end: the link only exists because a human being decided your resource was good enough to replace a dead one. That editorial decision is what makes it worth doing.
Step 1: find broken links that actually matter
You can find dead links in three ways, and I use all three depending on the campaign size.
Method A — crawl your competitors' resource pages
This is the workhorse. Take the top 5–10 sites ranking for your topic, pull their "resources," "links," or "further reading" pages, and run them through a broken link checker. Screaming Frog does this natively, and there are free browser extensions that handle single pages.
What you want in the output is: the page URL, the dead link URL, and the anchor text. The anchor text tells you exactly what the webmaster originally wanted to link to — which is your brief for what to offer as a replacement.
Method B — search for the dead resource directly
If you know a resource that used to exist (a tool, a study, a template that got taken down), search for its title or URL slug. You'll find pages still linking to it. Those are warm prospects because the anchor text is almost always descriptive.
Method C — AI-assisted prospecting
This is the newer piece, and it's the one I'd add to any workflow in 2026. I feed a list of candidate URLs to a model and ask it to flag which ones touch my topic and which ones look abandoned (last updated years ago, no author, no comments). It's not perfect. It's faster. I'd estimate it cut my manual qualification time by 40–50% on the last campaign.
What it does not do is improve reply rates. Every email still needs to sound like a person wrote it.
Step 2: qualify the opportunity before you write a word
The single biggest reason broken link building campaigns underperform is that people skip this step. Not every broken link is an opportunity.
What makes a link worth pursuing
- The hosting page is topically relevant to yours — not just vaguely adjacent.
- The page has some authority — it's linked to from elsewhere, it ranks, it gets traffic.
- The page is actively maintained — there's a recent update, a recent post, a live contact form.
- You have or can build a genuinely better replacement than what died.
- The dead link's anchor text is descriptive enough that your replacement fits it.
If any of these fail, drop it. I learned this the hard way: my early campaigns included pages with broken links from 2014 on sites that hadn't been touched since 2015. Nobody was answering those emails. I burned hours on prospects that were never viable.
The "no replacement" dead end
Sometimes the dead link pointed to something that genuinely has no modern equivalent. A discontinued tool. A dead government dataset. A blog that got taken down entirely. Move on. You cannot invent a substitute for a resource that no longer has a category.
Step 3: build or pick your replacement resource
Here's where most guides get vague. The replacement has to actually be good enough that a webmaster would have linked to it originally. Not "good enough given it's free." Good enough that if it had existed at the time, they'd have picked it instead.
Two options:
- Point to an existing page on your site that matches the anchor text. Fastest option, and if you already have a strong page, this is the play.
- Update an existing page so it now answers the dead link's intent. Slower, but you control the fit.
What I would not do is write a brand new page purely to fill this slot. That page will feel thin, and webmasters can smell it.
Step 4: write outreach that gets read
Keep it short. Under 100 words is the target. Mine usually land around 70.
The structure I use:
- One line explaining you found a broken link on a specific page.
- One line naming the dead URL.
- One line offering your replacement, with the URL.
- One line closing — no pressure, no follow-up promise.
No compliments about their content. No "I'm a big fan of your work." No paragraphs about who you are. The link report is the credibility.
Personalization that isn't performative
Personalization means the email is specific to their page. It doesn't mean you read their last blog post. I've tested both approaches. Specific-to-page beats specific-to-person almost every time in this tactic, because it directly references the problem you're asking them to fix.
Step 5: track, follow up, and learn
I send one follow-up, seven days after the first email, and that's it. Nothing more. My reply rates on follow-ups are low, and the marginal gain isn't worth the reputational cost of chasing.
Track:
- Prospects identified
- Emails sent
- Replies
- Links secured
- Links lost (the editor said yes but never updated)
That last one is real. Roughly a quarter of my "yes" responses never result in an actual link going live. Plan for it. A polite check-in two weeks later recovers some of those.
The tools you actually need
You don't need a paid suite to start. A crawler (Screaming Frog free version handles 500 URLs), a browser extension for spot checks, and a spreadsheet. If you scale past a few hundred prospects, a dedicated broken link checker or an outreach platform starts paying for itself. I've used both and honestly the spreadsheet was fine until I crossed roughly 400 prospects in a single campaign.
The part nobody tells you
Broken link building step by step is not complicated. The steps are: find, qualify, prepare, pitch, track. What's hard is resisting the urge to skip qualification because you're excited about a big domain. Every campaign I've run where I rushed that step underperformed. Every campaign where I built a tight list of 30 well-qualified prospects beat a loose list of 200.
The tactic hasn't changed much in a decade, and that's actually the point. While everyone else chases whatever link channel is trending this quarter, dead links keep accumulating on pages that still hold value. The web is quietly rotting in specific, findable places. Somebody has to point it out. Might as well be you.