Working out what a domain is actually worth

Read the three sale contexts, rank the evidence behind a comparable, and set a walk-away number before the conversation starts.

By Greg Priday · Published 3 February 2026 · Updated 16 August 2026 · 10 min read

Why an asking price tells you nothing

Domain pricing has no order book. There is no ticker, no bid-ask spread, and no obligation on a seller to price anything rationally. A name listed at $5,000 might have sold twice before at $400, or might genuinely be worth $20,000 to the one company whose brand it matches.

So the first move in any domain negotiation is not to react to the asking price. It is to build your own number independently, and only then look at theirs.

A worked example

Suppose a seller asks $5,000 for harborlane.com. A HumbleWorth estimate would return three figures, because there are three routes a name like this can sell through. Say they came back like this:

Example figures

Channel	Estimate	Likely range	What it assumes
Auction	$620	$285 – $955	Another investor buys it to resell
Marketplace	$1,800	$924 – $2,680	Listed publicly, waiting for an end buyer
Brokerage	$4,400	$2,460 – $6,340	A broker approaches a specific company

An example for an invented name, not a stored estimate. The estimates illustrate the shape of a report; the ranges around them are calculated exactly as the real report calculates them.

The $5,000 ask is not absurd. It sits inside the brokerage range. But brokerage is the ceiling scenario: it assumes a motivated corporate buyer and a broker doing outbound work. You are buying it on the open market, which is the marketplace scenario at best.

The relevant number is $1,800, not $5,000. That is the gap the negotiation is actually about.

Rank the evidence before comparing it

A model estimate is a starting point, not evidence. What moves a seller is a comparable. But "a real name with a real price" is not specific enough any more, because there are at least four kinds of number in circulation and they make very different claims.

What each number proves

01
Confirmed sale

Strongest

A transfer was reported at a price.

Reporting rarely says what else was in the deal, or who the buyer was.
02
Observed auction close

Strong

Bidding actually reached this level.

A reserve may have been missed, and an auction can be cancelled or the name reclaimed.
03
Asking price

Weak

A seller wants this.

No evidence any buyer agrees. Nobody has bid against it.
04
Model estimate

Not evidence

A modelled scenario for a name of this shape.

Not a transaction and not buyer intent. It is a starting point for the argument.

Know which rung your number is on before you put it in front of anybody, because the seller will. Quoting an asking price as market value, or an estimate as an appraisal, loses you the argument and some of your credibility with it.

The most useful practical change here is that closed auctions are now a source you can check. They are second-rung evidence rather than first, and they are public. Reading an auction result covers what that number does and does not prove.

Finding the comparable

Good comparables share as many of these as possible:

Same extension. .com comparables do not price a .io.
Similar word count and roughly the same syllable count.
Same category of meaning: two-word brandable, geo-service, dictionary word, acronym.
Same sale channel. An investor auction is not a brokered sale to an end buyer.
A similar market period, or an explicit adjustment for the time between.
A similar history profile: did that name carry a website, and does yours?
A similar traffic and backlink position, if either is part of the argument.
Known evidence quality, from the ladder above.

There is no hard rule that a comparable has to be from the last two or three years. Prefer recent evidence, but an older exact structural match adjusted for the market since beats a recent loose one that shares nothing but a word count.

One strong comparable beats five weak ones. If you can only find weak comparables, that is itself information: thin comparable data means a wide range, and a wide range means you should be anchoring low.

The comparable that is not one

The failure that does the most damage is a comparable that matches on every visible attribute and none of the invisible ones.

A weak two-word name closes at $8,000. Structurally it looks just like the name you are pricing. Then the archive shows a decade-old industry publication behind it, with real citations and traffic that still arrives. That result is evidence about a website and a domain sold together. It is not an $8,000 comparable for an unused two-word name, and using it as one will cost you the difference.

The inverse is just as common and gets noticed less. A strong name closes cheaply after a long spam history, and that result carries a risk discount somebody priced deliberately. Letting it anchor every clean comparable downward is the same error in the other direction.

So the archive check belongs inside the comparable process, not after it. Checking what a domain was before takes about two minutes per name and it is the step that decides whether a number is usable at all.

Browse recent closed auctions How to read those results

Setting a walk-away number

Decide this before you send the first message, and write it down. Once a conversation is running, every counter-offer will feel like progress, and progress is exactly what makes people overpay.

A simple rule that holds up well:

Opening offer	Around the auction estimate. Low, but not insulting, and defensible if challenged.
Target	The marketplace estimate, or a little under it.
Walk-away	The top of the marketplace range. Above that, you are paying brokerage prices on the open market.

The exception is when you are the end buyer: the name matches a business you already run. Then the brokerage figure is a fair ceiling, because you are the buyer the brokerage scenario assumes. Just be honest with yourself about which of the two you are.

Common mistakes

Treating one estimate as the price. A single number hides the spread, and the spread is where the negotiation happens.
Negotiating against the ask. Anchoring on the seller's number cedes the frame before you have said anything.
Mixing rungs of the evidence ladder. An observed close and a confirmed sale are not the same claim, and neither is an asking price.
Pricing a name against a website. Check the history of any comparable before you lean on it.
Ignoring renewal cost. A speculative buy at $1,800 also commits you to a renewal every year until it sells.
Skipping the trademark check. No estimate model assesses trademark exposure, and a cheap name with a legal problem is not cheap.
Buying on urgency. "Another buyer is interested" is unfalsifiable and free to say.

Closing safely

Once a price is agreed, the transaction risk shifts from estimating to settlement. Do not wire funds directly to a stranger against a promise of transfer.

None of this makes an estimate certain. It makes it defensible, which is the part that actually changes the price you pay. Run an estimate on the name you're considering and start from the marketplace figure.