What is a domain name worth?

A profit multiple is a price divided by a year of earnings. Most domains change hands with no earnings to divide by — no site, no revenue, no customers, just the registration and whatever a buyer thinks the name is worth to them. When a domain sells for six figures with no business attached, there is no multiple. There is a price.

So the figures above are deliberately lopsided. We track thousands of domain sales and publish no median multiple for them, because computing one would require a profit figure the sales do not state. Where a domain does come with an operating site, that deal is usually recorded under the asset type of the business — content, e-commerce, SaaS — and shows up in those medians instead.

This one catches almost everybody, and it changes how you read every domain price chart on the internet. The trade publications that report domain sales generally date a sale by when its price became public, not by when the transaction closed. DNJournal — a source we collect from, and the longest-running public record of domain sales — states this outright in its report on the $70,000,000 AI.com sale: listings on its charts have, in its words, "from day one 23 years ago, been based on when the price was first revealed - not the transaction date."

That sale is the example the publication itself reaches for, because the transaction had closed the previous spring. So a domain "sold in March" may have been agreed months or years earlier, and a chart of monthly domain prices is a chart of monthly disclosures. Reading it as a time series of market conditions will mislead you about when demand actually moved.

We keep that distinction rather than smoothing it away. Every comp we hold carries a sold_date_basis beside its sold_date, derived from the source the row came from, and rows from a reveal-dated source say reveal rather than close. It travels with the row into the Pro comps export, so a subscriber pricing a deal in a spreadsheet can see which dates mean what instead of assuming they are all the same kind of date. Where a report contradicts itself about its own sale, we store no date at all rather than a wrong one.

Every dataset of domain prices, ours included, is built from the sales somebody chose to announce. The publication we collect from is direct about what that excludes, writing in the same report that when it began verifying and charting sales in 2003 it knew "only a small fraction of overall sales would ever be publicly reported (especially at the high end of the market where non-disclosure is the usual standard)."

Two consequences worth carrying into a negotiation. First, the visible market is thinner than the real one, so the absence of a comparable sale for a name like yours is weak evidence about its value. Second, the sales that do get reported skew toward the ones worth reporting: a five-figure or six-figure name makes a weekly chart, and a $300 name usually does not. A quoted "average reported domain sale" is an average of what got written up, and pricing an ordinary name against it will overshoot badly.

With no multiple available, a domain valuation is a comparable-sales exercise and nothing else. The questions that actually move the number, in rough order of how much:

• Who the buyer is. The same name is worth one thing to an operating company that needs it and another to a reseller who has to find that company first. End-user prices and wholesale prices are two different markets, and most public reporting mixes them.
• Comparable names, not comparable industries. The evidence for a one-word .com is other one-word .com sales — length, extension, and whether the term is a real word someone searches for. A sale in the same industry with a very different name shape is not a comp.
• The extension. A .com and the same string on another extension are not interchangeable assets, and prices are not transferable between them.
• Whether anything comes with it. Traffic, an email list, backlinks, or a live site turn the deal into a business sale, which can be priced on earnings — and should be, using the comps for that asset type rather than domain sale prices.
• Carrying cost and time to sell. A name that is worth $20,000 to one buyer who has not appeared yet is not worth $20,000 today. Renewal fees are certain; the buyer is not.

None of those are numbers we can look up for you, and we would rather say so than dress a guess as data. What we can tell you is what names actually changed hands for and on what date basis, which is the raw material the exercise needs.

Domain deals fail on transfer mechanics more often than on price. Before agreeing a number, establish which registrar the name sits at, that the seller controls the account rather than a developer or agency who registered it for them, and that the name is unlocked with an authorization code available at closing. Registrar transfer locks can follow a change of registrant details, and the applicable window is set by current ICANN transfer policy and your registrar's implementation of it — confirm the timeline with the registrar before you commit to one, rather than assuming a figure you read somewhere still applies.

Then check what you are buying beyond the string: whether any trademark makes the name risky to use in your field, and whether the name has a history that follows it. Money should move through escrow against the transfer, not ahead of it. Our diligence checklist for sub-$100k deals covers the equivalent questions for deals that come with a business attached.