How to check what a domain was before

The archive, the backlinks, the ownership trail and the trademark check that separate a valuable name from a valuable old website, or from a liability.

By Greg Priday · Published 16 August 2026 · 13 min read

The name and the website are two assets

A domain is a registration. A website is content, code, links, traffic, a reputation, and sometimes a business. They share a name and almost nothing else, and buying one does not buy you the other.

Most of the expensive mistakes in aftermarket buying come from collapsing the two. Someone pays a website price for a bare registration because the auction ran hot. Someone else walks away from a good name because it once carried something ugly they could have cleaned up in a week. Both were reading a single price without asking what was being paid for.

So the question this guide answers is not "is this domain good". It is: what was here before, is any of it worth something to me, and what does the rest of it cost?

Three kinds of value

A domain can be valuable for three very different reasons, and a price on its own does not say which. Separating them is the difference between a comparable and a coincidence.

01

The name itself

The words, the extension, length, structure, how it sounds said out loud, and who would buy it. This is the part a valuation model is built to see.
02

A former website

Traffic, backlinks, citations, search visibility, or a business people already recognise. Real value, and none of it is in the string.
03

What that history costs

Spam, malware, adult or gambling use, topic churn, a manipulated link profile, or a trademark. Worth a discount, and sometimes worth walking away from.

No captures does not mean no history. The Internet Archive misses sites that were never crawled, were blocked or password protected, or were removed later. An empty calendar means the archive could not answer the question, which is a different finding from a clean one.

Start at the capture calendar

The Internet Archive is free, it needs no account, and it answers most of this question in about two minutes. Start there, every time, before you pay for anything.

Open the calendar, not a single snapshot. One capture tells you what a page looked like on one day. The calendar tells you how long the site ran, whether it ran continuously, and roughly when things changed, which is the part that actually informs a bid.

Then open three captures, not one:

the earliest one with real content on it;
one from the middle of the longest run;
the most recent one before the domain went quiet.

Three points is enough to see a trajectory. One point is enough to be misled by whichever day you happened to land on.

Reading the shape of a history

Before reading a single page, the calendar alone tells you a lot. Look at the distribution rather than the total.

What you see	What it usually means
Dense captures across many years	A site people visited and linked to. Crawlers return to pages that matter.
A long gap, then captures resume	The domain probably lapsed and was re-registered. Treat the two runs as two separate histories.
Thin, irregular captures throughout	A small or largely unlinked site. Not damning, but it caps what the history is worth.
Heavy captures that stop abruptly	A business that closed, or a site that moved. Worth finding out which.
Captures that are all parking pages	An investor held it. The name has no inherited website value either way.
Nothing at all	Unknown. Not clean. See below.

What a good history looks like

The signals worth something, roughly in order:

A single, legitimate site running for several years under consistent branding.
A topic that matches what the name says it is.
Evidence of a real organisation: an address, named people, a company number, a masthead.
Citations and mentions that were earned rather than placed.
Direct-navigation traffic that would plausibly still arrive.
No unexplained changes of subject, language, or ownership.

Be careful about how much of this you count on. Inherited search visibility is conditional on the new site being about roughly the same thing, and it decays. Rankings built by a trade publication do not transfer to a SaaS landing page on the same domain, whatever the backlink count says. Treat inherited website value as a reason the name might be worth more than the string, not as an asset you can bank at full price.

Red flags, and what they cost

Some of these are a discount. Some are a reason to stop.

Flag	Why it matters	Weight
Several unrelated topics in quick succession	Churned between owners or used as a disposable asset.	Discount
Mass-generated or doorway pages	Built to manipulate search, and search knows.	Heavy discount
Hacked pages or malware periods	May sit on blocklists that outlive the ownership change.	Heavy discount
Adult, gambling, pharma, counterfeit or piracy use	Filtering, payment-processor problems, and a brand association you cannot delete from people's memories.	Heavy discount
Link-network or expired-domain-network use	A backlink profile built to be sold, not earned. The count means nothing.	Heavy discount
Brand impersonation or a fake storefront	Complaints, chargebacks and possibly a dispute attached to the name.	Stop
A history dominated by someone else's trademark	The strongest reason a cheap name is cheap. Legal exposure is not priced into any estimate, including ours.	Stop
A high backlink count with no content history	Links from nowhere, for nothing. Usually purchased, sometimes hostile.	Investigate

"Stop" does not mean the name is worthless to everybody. It means the risk is legal or reputational rather than financial, and those are the two kinds you cannot price with a spreadsheet.

What no captures actually means

It means the archive could not answer the question. That is all it means, and it is worth being strict about, because "no history found" quietly becomes "clean history" in the space of about ten seconds.

The Internet Archive itself is clear that a site can be missing for reasons that have nothing to do with whether it existed: crawlers never discovered it, it was behind a password, it was blocked or excluded, it was unreachable when they tried, or it was removed afterwards. A site can also have been perfectly real, perfectly ordinary, and simply too small to be crawled.

Before recording an empty result, check the obvious causes:

try both www. and the bare domain;
try http as well as https;
search the name in quotes and see whether anything references a site that should have been captured.

Then write it down as unknown, and decide separately whether the bid is large enough to justify finding out properly.

Past the archive

The archive shows what a site looked like. It does not show who linked to it, whether those links carry any weight, whether the domain is currently indexed, or who owned it. For a material purchase you want at least three of the following.

Backlinks and referring domains. Not the count, the quality. A thousand links from sixty sites nobody has heard of is a red flag rather than an asset.
Index and branded search. Search the exact domain, then the brand name. Silence where a real business should be is worth a second look.
Ownership history. Historical and reverse Whois where it is available and lawful to use, mostly to answer how many hands the name has been through.
Trademarks. Search the relevant registers before you bid, not after. This is cheap, and it is the check people skip most often.
Reputation and blocklists. Security and spam lists, and a look at whether mail from the domain would even deliver.
Claimed traffic. If a seller says the name gets visitors, the burden of proof is theirs, and analytics screenshots are not proof.

What does not transfer with the domain

Worth being explicit, because auction listings never are. Winning a domain auction normally transfers the registration. Unless a contract says otherwise, it does not transfer:

the previous site's content, or any right to republish it;
its code, design, or anything licensed to the former owner;
customer data, mailing lists, or accounts;
social profiles or handles under the same brand;
trademarks, or any right to trade under the old name;
contracts, listings, or business relationships.

Archived pages are a record of what was published, not a licence to publish it again. If your plan for a name depends on restoring the old site, that is a plan that needs a lawyer rather than a backorder.

There is one more consequence people miss: if the previous owner's brand was genuinely well known, inheriting their traffic can mean inheriting their customers' expectations too. That is worth something in some businesses and is a support problem in others.

Write the verdict down

Research you did not record is research you will do again. Keep a note per name, in the same shape every time, so two names can be compared later without reopening either.

Research note template

Domain:
Archive coverage:      strong / partial / none visible
Earliest capture:
Latest capture:
Primary historic use:
Topic continuity:      stable / changed / highly unstable
Positive signals:
Liabilities / flags:
Included in purchase:
Still unknown:
Effect on value:       increase / neutral / discount / reject
Reason:

One deliberate omission: there is no clean on that list. It is a word that means "I did not find anything", dressed up as a finding. Use one of these instead:

Stable and relevant history. Long single use, matching topic, real signals.
Useful, but it is the website's value. Price the name and the site separately.
Mixed history. Bid with a discount that covers the cleanup.
High risk. Do not proceed without a closer look, and possibly not then.
Unknown. Not enough evidence either way, and the bid should reflect that.

A due-diligence pass, by budget

Match the effort to the bid. Doing fifteen minutes of work on a $30 name is how people stop doing any research at all.

Effort	What you do	Worth it when
2 minutes	Capture calendar, three captures, one-line verdict.	Always. Including hand registrations.
5 minutes	Add a branded search, a backlink summary, and a trademark search.	Anything you would be annoyed to lose money on.
15 minutes	Referring-domain quality, ownership changes, reputation and blocklist checks.	A bid that is material to you, or a name with a website thesis attached.
Proper review	Legal advice on the trademark position, technical review of the link profile.	A purchase large enough that being wrong changes your year.

The times are targets for how the work is scoped, not promises about how long it takes.

None of this makes a name safe. It makes the decision an informed one, which is the most any research gets you, and it is the difference between a discount you chose and a discount you discovered afterwards.

Take a name you are considering, run the estimate, then run the two-minute pass above. If those two disagree, the history is usually the one telling the truth.

The limits described under “What no captures actually means” follow the Internet Archive's own guidance on why a site may be absent from the Wayback Machine.