Transcript
Intro
A dental clinic in Istanbul came to us with a site that looked fine and performed like it was hiding. Machine translated pages in eleven languages, every one on a query string address, and a backlink profile full of spam. On the twenty seventh of May, before we started, it got eight clicks from Google. On the seventeenth of September, it got fifty eight.
We'll also meet a patent firm whose site was rebuilt by another team without a single redirect. Both got their traffic back. But under every rescue like these sits one decision: repair the site where it stands, or move it to a new domain. And if you move, how you move decides whether the damage comes along.
To keep the theory clean, we'll use ReelRadar, our fictional film trailer site, after a few years of bad advice. One agency bought watch movies online anchors from sites in languages nobody there speaks. Another ran automated campaigns from casino and pharma pages. The content's fine. The link profile reads like a spam directory.
One warning first. All of this assumes the links really are the problem. Plenty of sites blame their backlinks for a drop that came from thin content, a botched migration or a core update. So step one is a proper diagnosis.
Diagnose, then repair
There are two kinds of link hit. A manual action is a person at Google deciding your site broke the link spam policy. It shows up in Search Console's Manual actions report as unnatural links to your site, and you get a message. You know it happened, and you know why.
The algorithmic kind is quieter. No message, nothing in the report. Visibility just sinks, often around a confirmed spam or core update, and the pages with the worst links fall hardest. Google has said since 2016 that its link spam system devalues spam signals, so you're reading signals, not notices.
Run three checks. Open the Manual actions report, which settles the first question in seconds. Line the drop up against the update dates on Google's Search Status Dashboard. Then split your pages by how bad their links are and compare. If clean pages held and poisoned ones sank, the links are your likely culprit.
Is a competitor attacking you? Rarely, in our experience. Google says it works very hard to stop third party actions from harming a site, and spam you didn't ask for is what the neutralising systems were built for. A truly buried site has almost always been buried by links somebody paid for.
Option one is repairing the site where it stands. For a manual action, Google's process is clear: remove the links you can, disavow the ones you can't, and file a reconsideration request. Google's disavow page describes this exact case, including links you, or an SEO you hired, built through paid schemes.
A few practical notes from that page. Disavow single URLs, or whole domains with the domain prefix, one per line. The tool doesn't support Domain properties, so you need a URL prefix property. And a new upload replaces your old list, so keep one master file and add to it.
The reconsideration request is where most people underperform. Google wants evidence of work, not an apology. Say what happened, show what you removed and how you tried with the rest, attach the disavow file, and explain what stops it happening again.
For an algorithmic hit there's nothing to file. Repair means changing the signals. Disavow the links you know you built, then dilute the anchor pattern with relevant links carrying natural anchors, as in the anchor text episode. Once the bad pattern is a minority, recovery becomes possible.
Claim number one. Disavow everything and wait, and the algorithm will let you back in.
For an algorithmic hit, that's a placebo, and sometimes worse. Google says the tool is for links that caused, or will likely cause, a manual action, and that used wrongly it can hurt your site. Disavowing everything also throws out the good links holding you up. And waiting changes nothing, because the pattern Google reads is still sitting there.
Repair is the right call when the damage is limited: a slice of bad links, a topic that still matches, an anchor pattern you can realistically outweigh. It keeps your domain, your history and every good link. Changes only count once Google recrawls the linking pages, so give it a fair run before you judge it.
Rescue 1: a dental clinic
Here's a repair in place on a real site. The dental clinic serves patients who travel from abroad, so it ran eleven languages. Every one was machine translated, and nobody had checked them. Every language lived on a query string address, the kind that ends in question mark, lang equals de. And the backlink profile carried a lot of spam.
Google's own guide to multilingual sites lists language URL parameters as not recommended. It wants a distinct URL for each language version, with hreflang tying the versions together. The clinic had the opposite on every count, and its clicks had sat flat for months.
So we fixed it where it stood. No new domain needed. Proper translations replaced the machine ones. Every language moved to a clean URL, with a 301 from every old query string address to its own new page, plus hreflang. And we cleaned up the link profile. The domain was worth keeping, because its problems could be fixed in place.
The result, straight from Search Console. Before us the clinic averaged nineteen point six clicks a day. September averaged thirty four point eight a day, up seventy seven percent, and daily impressions were up a hundred and twenty three percent. That's the whole repair at work: proper translations, clean addresses and a cleaned-up link profile.
Moving to a new domain
Now option two, for when repair can't win. Thousands of links carrying irrelevant terms or exact match money phrases, from sources you could never remove, in a pattern no realistic amount of natural linking would outweigh. At that point you're paying to fight your own history.
So you move to a new domain and take the business with you: the content, improved where it needs it, the brand, the customers and their trust. What stays behind is the link profile that buried you. That fresh start is the whole reason to move.
Pick the new domain like an asset, not a throwaway. Keep it close to the brand so customers recognise it. And if it has any history, check it first: Google's site move guide tells you to look for manual actions left behind by a previous owner.
Treat it as a business decision too. A new domain means a new address on invoices, ads, listings and packaging, and some customers will be confused for a while. Tell customers and real partners yourself, and ask the people who linked to you honestly to point at the new address. Those are the cleanest links you'll ever get.
And here's the trap. The obvious way to move is a straight 301 from the old domain to the new one. That's how a normal site move works, and a normal move is designed to carry everything across. With a poisoned profile, everything includes the poison.
Google's documentation is plain about what redirects do. Keep them for as long as possible, it says, so Google can transfer all signals to the new URLs, including links from other sites. All signals means the bad ones too.
Google has even said it out loud for one case. In its 2024 site reputation FAQ: if you move content that got a manual action, don't redirect the old URLs to the new site, because that may bring the issue back. That FAQ is about parasite content, not link spam, so don't stretch it. But the principle carries. A redirect is a pipe, and Google follows pipes.
Claim number two. Just 301 the old site to a new domain and start fresh.
Contested, and risky. For an ordinary migration with a clean profile, a straight redirect is exactly right. For a site buried by its links, it's the move most likely to bring the burial along. Some sites get away with it. Many watch the new domain sink for the same reasons the old one did.
The bridge-domain method
So, the bridge domain, as a principle. Old domain A doesn't point straight at new domain B. A redirects to a bridge domain, C, and C redirects on to B. In the cases where we've used it, much of the toxic effect stayed stuck in A and C, while a useful share of the link power still came through to B.
And it isn't one big redirect to a homepage. Every page goes to its own new address. ReelRadar's horror streaming page on A goes to the same path on C, and C sends it on to the same page on B. Redirect rules map the paths automatically, so the whole site moves page for page, without anyone matching URLs by hand.
Why would a middle step help? Here's our working explanation, labelled as ours, because Google has never described this. Each hop is a point where signals get evaluated again rather than simply passed along. And C is a control point. If it starts to look like it's carrying the problem, you can cut it, and B was never directly tied to A.
The bridge has to be clean. A bridge with its own spam history, or one stuffed with throwaway content, filters nothing. It just adds a second problem on top of the first. Treat it as a credible domain in its own right, not a pipe with a name on it.
What we won't tell you is just as deliberate. No counts, no timings, no hosting details, no setup. The method works because each of those choices is made for the site in front of us. Copied as a recipe, it creates exactly the footprint it's meant to avoid.
Claim number three. The bridge domain move can put a buried business back on its feet.
Our verdict: it works, when it's done precisely. We've seen it protect the new domain where a direct redirect would have carried the damage across. That's our field work, and the setup stays private, because the details are exactly where most attempts go wrong.
Rescue 2: a patent firm
Now the opposite problem: no redirects at all. A patent and trademark firm, in business since 1999, had its site rebuilt from ASP.NET to WordPress by another team. Nobody set up a single 301. Every old address went to a 404 page, every link pointing at those addresses hit a dead end, and organic traffic slid for a long time.
We rebuilt it on a custom Next.js stack and redirected every old URL, from the ASP.NET era and the WordPress era, to the right new page. Not to the homepage. To the page that actually replaced it. We also restored the English pages, fixed hreflang and structured data, and took mobile Lighthouse scores from the sixties to above ninety.
The result: from about four hundred organic visitors a month to nearly four thousand, in under three months. Up eight hundred and thirty percent. The links the firm had earned over more than twenty years had been pointing at dead pages. Once each one led somewhere real again, that value came home.
Notice what the two rescues share with the bridge. It's all redirect work done page for page. The clinic's old language addresses each went to their own new page. The patent firm's dead addresses each went to their true replacement. The bridge does the same thing, with one extra hop, for a domain you need to leave behind.
Costs and failure modes
Now the honest costs of a move, because none of this is free. Some good link power gets lost with the bad. A move takes months, not weeks, because Google has to recrawl and re-evaluate everything. And the new domain still needs fresh, clean links of its own. The bridge protects the move. It doesn't replace building authority.
Measure it honestly. Verify the new domain in Search Console on day one. Track impressions first, then clicks, then position on the queries that matter. And compare against where you were before the hit, not against the lowest point. Recovering from the bottom is easy to claim. Getting back to where you were is the real test.
Build the new domain's own links from the start, the boring way: brand and address anchors, real partners, real coverage. A clean, brand led profile makes the move safer every month. Repeat the old pattern and you've just bought a second domain to bury.
Most moves fail for predictable reasons. People move when repair would have worked, and lose a good domain for nothing. They send every old page to the new homepage instead of to its own counterpart. They use a bridge that's a spam site itself. Or they rebuild the same bad links within months.
And one that really hurts. If the old site has a manual action for its content rather than its links, moving doesn't wash it away. Google's guidance for moved content is not to redirect it, and to handle the manual action properly. Know which problem you have before you pick the tool.
Run the decision on ReelRadar. No manual action, a drop after a spam update, and the casino linked pages fell hardest while the clean pages held. The profile is dominated by links nobody can remove, so dilution is out of reach. The answer is to move, through a clean bridge, with every page going to its own new address.
That's why we call it precision work, not a weekend job. Which domain, what the bridge looks like, when anything happens: every choice can tip the result. If the business can't afford a mistake, get someone who's done it before, or stay with repair.
Wrap up
The episode on one screen. Disavow is for the cases Google describes. A straight redirect is a gamble with a poisoned profile. Redirects done page for page rescue sites, and the bridge works when it's done precisely.
The decision in one line. Diagnose first, repair whenever the damage is limited, and move only when the profile is beyond dilution. If you move, never pipe the poison straight into your new home.
The repair or move decision sheet is free on the episode page at irongoo.com. Next: the advanced methods, tiered links, private networks and paid placements, and why most people who try them fail.
