How to Evaluate an Expired Domain's Link History

The first expired domain I got badly wrong had 412 referring domains. I remember the number because I kept looking at it, the way you look at a price tag that seems too good.

It was worthless. Nearly all 412 were directory listings and scraper sites, the same footer link repeated across a network that had clearly been built and abandoned by one person. Two of the references were real. Two. The rest were noise that no search engine had cared about since about 2013.

That is the whole problem with evaluating link history. The number that is easiest to find is the number that tells you the least.

Start with who, not how many

A referring-domain count answers one question: how many separate sites point here. It says nothing about whether any of those sites matter, and the distribution is always brutally uneven. On most domains worth having, three or four references are carrying the entire value and everything else is filler.

So the first thing to look at is the list of linking domains, read as a list of names. Not scored, not averaged, just read. You are looking for sites you recognise, or sites you can check in ten seconds and form an opinion about. A university library page. A government agency's resource list. A trade publication that still employs editors. A reference work.

One link from a source like that is worth more than two hundred from anywhere else, and you do not need a tool to tell you so. You need the names.

This is also the check that separates a real history from a manufactured one. Manufactured link profiles look identical in aggregate and fall apart on inspection: the same handful of hosting patterns, the same site templates, article pages with no author and no other outbound links, domains whose own history begins and ends with link selling. Once you have seen a few, the pattern is hard to unsee.

Editorial or placed?

The question behind the question is whether someone chose to link, or whether the link was bought, submitted, or automated.

A few signals separate them. Editorial links sit inside prose, on a page that exists for a reason other than linking. They point at something specific: a page, a document, a dataset, not always the homepage. They have varied anchor text, because whoever wrote the sentence was writing a sentence, not optimising an anchor. The page around them has other outbound links to unrelated places.

Placed links do the opposite. They cluster in footers, sidebars, author bios and resource lists. They use exact-match commercial anchors. They arrive in bursts, dozens in a month, from sites with no other editorial connection to each other.

You can usually tell within a minute of opening three or four linking pages. Open them. This is the step people skip and then regret.

Does the page still exist, and does anyone maintain it?

A link is only worth what its host page is worth right now. Pages rot: they get redesigned, consolidated, moved behind logins, or quietly deleted. A reference from 2011 on a page nobody has touched since 2012 is real, but it is doing less work than a reference from 2019 on a page that gets updated every quarter.

Check the linking page loads. Check the link is still in the HTML rather than remembered by a stale index. Check whether the page looks maintained. Search engines respond to maintained content, so a well-kept old page can carry more weight than a neglected recent one.

What did the domain actually publish?

Now leave the links and look at the domain itself, through its archive history.

You are answering three questions. Did this domain have a real site, or was it parked? What was it about? And is there a stretch where it turned into something you would not want to inherit?

That last one matters most. Plenty of respectable domains spend two or three years, usually right before they lapse, hosting gambling affiliate pages or worse, and the links from the respectable era survive while the reputation from the bad era survives too. A continuous archive with a consistent subject is the good case. Gaps are normal. A subject that changes abruptly from a regional charity to online casinos is a domain to walk away from, whatever the referring-domain count says.

Google folded this into its published spam policies in March 2024 under the heading of expired domain abuse, and its definition turns on intent: an expired domain purchased and repurposed "primarily to manipulate Search rankings by hosting content that provides little to no value to users". Our guide to redirecting an aged domain goes into where that line sits in practice.

Ignore the composite score, mostly

Domain Authority, Domain Rating and Authority Score are third-party estimates. Google does not read them, and they can be moved with money: documented experiments have paid between $15 and $85 per domain to push a fresh domain's Domain Rating into the 40s and 50s within days.

That does not make them useless. A composite score is a fine way to sort a long list before you look at anything properly. It is a terrible way to make the final call, because the thing it measures is the thing that is easiest to fake. Domain authority explained covers what these scores are actually derived from.

Treat the score as a filter and the linking domains as the evidence.

Relevance, and being honest about it

An aged domain about veterinary equipment does not help a fintech site much, whatever its metrics. The links were earned in a context, and the context travels with them.

Relevance is a spectrum rather than a gate. Adjacent is usually fine. Unrelated is not fatal if the references are strong enough and you are building something real on the name rather than pointing it at an unrelated page and hoping. But a domain whose history is close to your subject is worth paying more attention to than a domain with better numbers and no connection to anything you do.

Is it even still available?

This one is embarrassing to get wrong and easy to check. Availability data goes stale fast, and a domain that was free when a list was compiled last Tuesday may not be free now. Look for a timestamp on the availability check, not just a status. If a directory will not tell you when it last looked, assume it has not looked recently.

The order I actually work in

Sort by whatever composite score is handy, because you have to start somewhere. Open the linking domains and read the names. Open three or four linking pages and decide whether a person or a script put the link there. Scan the archive history for a subject and for the years you would rather not inherit. Then, and only then, care about the count.

It takes about five minutes per domain once you have done it fifty times. The first ten will take twenty minutes each, and one of them will save you from a 412.

Every listing in the Revised directory publishes this evidence up front: the referring-domain band, the sources that link to it, the archived snapshot history, a clean-history screen and the date availability was last checked. The metrics are public without an account. Opening a domain's name is what a plan is for, and the free tier includes enough reveals to work through a shortlist before you decide whether the data is worth paying for.