What link reclamation is
Start with the links you own. When I ran a full audit of my own profile, two cross-links from billhartzer.com to hartzer.com were flagged as deleted — one on the phrase about buying brand mentions, one on an expert witness page. Both ends of both links are mine. Both had rotted within the previous three months, through nothing more dramatic than a page edit. Ten minutes of work, and it had not occurred to me to check.
That is the shape of the whole tactic. Link reclamation is recovering links you used to have and lost, plus fixing links that exist but do not work — pointing at redirect chains, at 404s on your own site, at old protocol or host variants, at pages you consolidated during a migration.
It divides into two halves. Internal reclamation repairs what is broken on your side: redirects, retired URLs, changed slugs. That half needs no outreach and should be done first. External reclamation is contacting publishers who removed a link and asking for it back, and that half is where the evidence gets interesting.
Why the prospects are alive and warm
Everyone believes links disappear because pages die. Sites shut down, posts get deleted, 404s accumulate. It is the standard explanation and it is almost entirely wrong.
Across 645,202 lost links on a single profile — 1,301,839 links from 22,260 referring domains, none of them ever bought — the recorded reason at the moment each link was lost breaks down like this: 78.0% were lost while the page fetched perfectly and simply no longer contained the link; 19.4% were lost to a canonical change; a further 2.5% to redirects. Dead hosts, DNS failures and timeouts together accounted for 24 links. And exactly 4 links out of 645,202 were lost to a 404.
97.4% of lost links were lost while the source page was still perfectly reachable.
Three consequences follow, and each one rewrites how you should work.
- The prospect is alive. The page still exists, still gets fetched, and in many cases still ranks. There is a live human maintaining it.
- The pitch is warm. This person linked to you once. That is a completely different conversation from broken link building, where you are a stranger reporting a fault.
- Monitoring 404s catches nothing. The standard advice — watch for dead referring pages — would have caught four links out of 645,202.
On that same profile, 717 referring domains at Trust Flow 21+ (Majestic's measure of how trustworthy the sites linking to a page are, on a 0-100 scale) have stopped linking entirely. Seven hundred and seventeen organizations that once judged the site worth citing, whose pages are almost certainly still live, and nobody has asked them why they stopped.
How it works, step by step
- Fix your own site first. Export every referring URL from a link index, crawl the target URLs, and find every inbound link pointing at a 404, a redirect chain, or a non-canonical host variant. Repair with direct 301s to the right destination. This recovers value with zero outreach.
- Check the links you control on both ends. Your other properties, partner sites, author bios, profiles you own. Mine had two dead ones I did not know about.
- Pull the lost link report. Every serious link index will give you links or referring domains recorded as lost, with the date last seen.
- Filter by quality and recency. Sort by Trust Flow or an equivalent, discard anything below your threshold, and prioritize losses from the last twelve months — a page edited recently is edited again.
- Verify the loss. Load the page. Is the link genuinely gone? Was the whole section removed? Was it replaced with a link to a competitor? The answer changes what you write.
- Diagnose before you write. A redesign, a migration, a policy change and a deliberate removal need four different emails.
- Contact the editor. Reference the specific article, note that it used to cite your page, and ask whether the removal was intentional. Often it was not.
What it costs in time and effort
The internal half is one focused day of a technical person plus a link index subscription you already have, and it is the highest-return day in link building. The external half costs about what any outreach costs, but the yield per hour is substantially better, because every prospect is pre-qualified by the fact that they linked to you once already.
The recurring cost is a monitoring habit. Link loss is continuous: on my profile the median referring domain stops linking after 1,080 days, and 49.6% of all links ever recorded have been lost. Standing still requires replacing roughly a sixth of the profile a year, and a quarterly reclamation pass is the cheapest way to do it.
When it works and when it does not
It works whenever a site has history. If you have been publishing for five years you have a reclamation list whether you have looked at it or not. It works especially well after a site migration, a rebrand, a domain change or a CMS replacement, because those events generate large, concentrated, entirely fixable losses.
It works when the removal was incidental — a redesign that dropped a sidebar, a template change, a bulk canonical update. Those get restored at a good rate because nobody decided anything.
It does not work when the link was removed on purpose. If an editor cut you because your page went stale, went behind a paywall, or started reading like an advertisement, fix the page rather than sending an email. It does not work on sites that have been sold or redirected wholesale. And it does not work for a site with no history: a business six months old has nothing to reclaim.
Common mistakes
- Watching for 404s. The most persistent piece of bad advice in this area. Four links out of 645,202. Watch for link removals on live pages instead.
- Skipping the internal fixes. Chasing publishers while your own redirects are broken is doing the hard half first.
- Treating a redirect as safe. A link into a redirect chain still passes value, but chains break during migrations and canonical changes are the second largest cause of loss in the data. Point the link at the live URL.
- Emailing before diagnosing. Asking why a link was removed when the whole page was rewritten makes you look like you did not read it.
- Reclaiming junk. On my profile 69.3% of referring domains are Trust Flow 0, and nothing was ever bought. Losing those is not worth an email.
- Doing it once. Reclamation is a quarterly habit, not a project. Loss is continuous; recovery has to be too.
A worked example
Take the 717 TF 21+ domains that stopped linking to my own site. That list is a working prospect file, and it runs in three passes.
Pass one is verification. Load each historic referring URL, confirm the page still resolves — on this evidence, it nearly always will — and record what replaced the link. You find four patterns: a redesign took the resource list with it; the article was rewritten and the citation dropped; the site migrated and canonicals now point elsewhere; or a competitor's link sits where yours was.
Pass two is triage. Redesign and migration losses go to the top, because those removals were not decisions and the editor is usually glad to hear about them. Deliberate replacements go to the bottom.
Pass three is the email, and it writes itself: your article on this subject cited my analysis of that subject until roughly this date, the reference appears to have come out during your redesign, and I wanted to check whether that was deliberate. No pitch, no anchor text request, no attachment.
The largest single item on my own list is a domain I own that was once the third-largest referring domain in the profile and now sends nothing at all. No outreach required. That is the point about starting with the links you own.
How to measure it
Measure referring domains recovered, not links, and measure them against the size of the loss pool so you know your recovery rate. Track internal fixes separately from external wins.
Then track the loss rate itself: referring domains lost per quarter, and the Trust Flow distribution of those losses. A rising loss rate at high Trust Flow is an early warning that matters far more than a flat total link count, and it is invisible on any dashboard reporting only totals.
Set a floor. Decide the quality threshold below which a lost link does not get chased, and stick to it, or the sheer volume of low-value loss will absorb all your time. Re-verify recovered links after six months.
The verdict
This is the tactic I would put first for any established site, and the evidence behind it is the strongest on this site. The received wisdom — links die because pages die — is wrong by a factor that is hard to overstate: four 404s out of 645,202 lost links. Everything else was a live page that quietly stopped mentioning you.
That converts link loss from an act of nature into a list of warm prospects. The pages still exist, still rank, and still have somebody maintaining them. You are not cold-emailing a stranger about a fault on their site; you are writing to a publisher who already decided, once, that you were worth citing.
Start with the links you own, because they cost nothing and I found two dead ones on my own properties without looking. Then fix your redirects. Then work the list of high-quality domains that stopped linking. Do it every quarter, forever. It is cheaper than acquisition, it is entirely safe, and almost nobody bothers.
Questions
How many lost links can I realistically recover?
There is no reliable published rate, and I will not invent one. What the evidence establishes is that the prospects are recoverable in principle: 97.4% of lost links on a 1.3 million link profile disappeared while the source page stayed live. Recovery depends on whether the removal was incidental or deliberate, and incidental removals — redesigns, migrations, template changes — convert far better.
Should I monitor 404s on referring pages to catch link loss?
No. On 645,202 lost links, exactly four were lost to a 404. Monitoring dead referring pages catches essentially nothing. Monitor lost links directly in a link index, and watch for links that vanish from pages which still fetch perfectly — that is 78% of all loss, plus another 19% to canonical changes.
Is link reclamation the same as broken link building?
No, and the difference is the whole point. Broken link building contacts a stranger about somebody else's dead resource. Reclamation contacts a publisher who linked to you and stopped, on a page that is still live. One is a cold pitch about a fault; the other is a warm follow-up with a shared history. The reclamation pitch is far stronger.
How often should I run a reclamation pass?
Quarterly for an established site, and immediately after any migration, rebrand or redesign. Loss is continuous rather than episodic — on my own profile the median referring domain stops linking after about 1,080 days and roughly half of all links ever recorded are gone. A once-a-year pass leaves nine months of loss to go cold.