The past does not transfer away
Buying a domain without checking its history is the most expensive routine mistake in this field, because the damage is invisible until after the money has moved and it does not announce itself.
A domain accumulates a record. Search engines remember what was published on it. Mail providers remember what was sent from it. Security vendors remember what was hosted on it. Trademark holders remember what was claimed on it. None of that resets when the registrant changes, and much of it is not visible from the current website, which may be a clean parking page sitting on top of a decade of something else entirely.
The asymmetry is what makes this worth systematizing. The check costs a small fraction of a typical purchase price and takes hours. Discovering afterward that a name has an inherited penalty, a blocklist entry that keeps your email out of inboxes, or a prior use that reads as trademark infringement costs the purchase price plus everything you built on top of it.
Why I built a process for this
In 2013 I created an algorithm and a patent-pending process for performing a background check on an internet domain name. The reason was mundane: this work was being done inconsistently, by hand, by people who each remembered a different subset of the checks, and the results were not comparable between one name and the next.
A repeatable process solves two problems. It stops checks being skipped because whoever ran them was in a hurry, and it produces output that can be compared across names, which is what anyone evaluating several candidates actually needs. That process became the basis of the domain background check services I have run since — first as Verified Domains in partnership with Globe Runner, then in automated form through DNProtect, which I co-founded with Rob Monster in 2020 and served as Director of, and now through DNAccess, which I founded in March 2023.
What follows is the shape of the examination rather than the scoring internals. Anyone can run most of it. The value is in running all of it, in the same order, every time.
Registration and lifecycle history
Start with the record, because it frames how you read everything else.
- Creation date, and whether it is the original one. A domain that dropped and was re-registered typically shows a new creation date while keeping the accumulated links and associations of its previous life. The gap between the earliest archived content and the stated creation date is the fastest way to spot this.
- Ownership changes over time, from historical registration archives assembled before contact redaction became standard and from the commercial providers that retained them. Frequent handovers, or a change that coincides with a dispute or an outage, both deserve explanation.
- Registrar history. Movement between registrars is normal. Movement to a registrar with a poor abuse reputation, at the same time as a content change, is not.
- Status codes and lifecycle events. Whether the name has ever been placed on hold by a registry, which is not something that happens for administrative reasons.
- Dispute history. Published administrative decisions are searchable by domain name, and a name that has already been the subject of a complaint tells you exactly what argument the next complainant will make.
Read all of this as a timeline rather than as a set of fields. The interesting findings are almost always about coincidence in timing.
What was published on it
Archived captures are the core of the examination, and most people run this step far too shallowly — they look at the homepage on a few dates, see nothing alarming, and stop.
Go deeper. Sample across the whole history rather than the recent past, because the problematic period is usually years back. Look at internal pages and not only the front page, since a site can present a legitimate face while running something else in a subdirectory. Check every language the site has operated in; a name that looks like a defunct local business in English may have spent three years as something entirely different in another language, and that is one of the most common surprises in this work.
Note the specific categories that follow a name: adult content, gambling, pharmaceuticals, counterfeit goods, cryptocurrency promotion, and anything that looks like a doorway page or a link farm. Note whether the site was ever a genuine business with real content, which is the profile you want, or whether it has been a rotating series of thin sites, which is the profile that suggests it was passed between people using it for the same purpose.
Also note what is missing. Long gaps with no captures, or a robots exclusion that removed the archive retroactively, are not proof of anything, but they are a reason to look harder elsewhere rather than to relax.
The inherited link profile
The links come with the name, and this is where the “aged domain with authority” pitch usually falls apart.
Pull the referring domains from more than one index and clean the export properly before judging it: normalize the URLs, collapse site-wide template links, roll subdomains up to registrable domains, and verify which links still exist on live pages. What remains is the actual profile, and it is normally a fraction of the advertised figure.
Then read its shape. A profile from a real business looks lumpy and topical, dominated by brand and bare-URL anchors, acquired over years. A manufactured profile shows commercial anchor text at the top of the distribution, bursts of acquisition confined to short windows, and referring domains that cluster on shared nameservers, shared analytics identifiers or adjacent address ranges. Foreign-language links with commercial anchors on a domain that never operated in that language are the clearest tell there is.
Judge relevance last and judge it honestly. Links from sites with no topical relationship to the name's proposed new use are not an asset you are acquiring; at best they are inert, and at worst they are a cleanup project you are paying a premium for.
Reputation, security and trademark exposure
Three further layers, each of which follows the name rather than the owner.
Reputation. Check spam blocklists, browser safe-browsing status, security vendor categorization, and mail sender reputation. Delisting is often possible but it is a process with its own timelines, and if the business depends on email from day one that timeline is your launch date. Content filtering categories matter too: a name categorized as adult or gambling by an enterprise filter will be blocked on corporate networks long after the content is gone.
Security history. Whether the name has appeared in malware or phishing feeds, whether it has ever hosted a command and control endpoint, and whether it appears in public breach or abuse datasets. This history is sticky and it is held by parties who will not notice your ownership change.
Trademark exposure. Search the registers in the markets you operate in, not just one. A name that reads as somebody else's mark is the fastest available route to a dispute you will lose, and the risk is not removed by intending to use it differently. Prior use of the name in a way that infringed is a separate and compounding problem, because it establishes exactly the pattern a complainant needs to describe.
Scoring, and what a score is not
A composite score is useful for triage — comparing forty candidate names, or flagging which of a portfolio deserve attention. It is not a verdict, and I would not want anyone to treat it as one.
The reason is that the same finding carries different weight depending on intent. Adult content history is close to fatal for a family brand and nearly irrelevant for a name that will only ever be a redirect. An inherited link profile built for a different industry is a liability for a site trying to rank and immaterial for a name being bought defensively. A score cannot know which of those you are.
So the deliverable that matters is the evidence behind the score: the dated captures, the ownership timeline, the cleaned link data, the blocklist entries with their dates. Anyone relying on this to make a decision — or later to explain a decision to somebody else — needs the underlying material, not a number.
What ends a deal, and what only changes the price
After enough of these, the findings sort into two piles.
Walk away from names with unresolvable trademark conflicts in your market, from names with a history of malware distribution or phishing, from names whose ownership history suggests a theft that was never resolved, and from names whose entire apparent value is a link profile that was manufactured. That last one is worth stating plainly because it is sold hardest: an expired domain marketed for its accumulated authority, with a profile built by whoever was farming it, is not an asset. Redirecting it at your site is a tactic search engines recognized and discounted years ago.
Negotiate on recoverable problems: blocklist entries that can be appealed, an unremarkable prior use in an unrelated field, a modest inherited profile that needs disavowing, or a lapse that reset the creation date without changing much else. These have a cost and a timeline, and both belong in the price.
And document the check regardless of the outcome. If the name later becomes the subject of a dispute, the fact that you examined its history before purchase and recorded what you found puts you in a materially better position than a buyer who did not look. Domain name background checks run through DNAccess.