Mastering Comparable Sales Data

Understanding mastering comparable sales is the foundation of this guide. Below we break it down step by step so you can act with confidence.

The difference between a professional domain trader and a hopeful amateur is almost never taste or instinct. It is pricing, and pricing comes from data. Amateurs look at a name they own, feel a number in their gut, and attach it with a mental story built around the one headline sale they read about last year. Professionals do the opposite. They start from what names like this one have actually sold for, in real closed transactions, recently and verifiably, and they let that record set the range before emotion ever enters the room. The instrument that makes this possible is the comparable sale, or comp, and the discipline of using comps is what separates a portfolio that turns over at a healthy sell-through rate from one that sits unsold for a decade while the renewal fees quietly bleed it dry.

This lesson walks through how comps really work: where verified sales data lives and what each source actually measures, how to find names that are genuinely comparable rather than merely similar-looking, how to adjust a comp up or down for the handful of traits that truly move price, why you weight recent and verified sales and throw out the outliers, and the wide gap between the asking prices plastered everywhere and the sold prices almost no one advertises. By the end you will be able to build your own private comp sheet, the single habit that stops you from both overpricing a name into permanent silence and underselling a good domain for a fraction of what the market would gladly have paid.

Mastering Comparable Sales: Where Real Sales Data Actually Lives

Most beginners price in a vacuum because they do not know the public record exists. It does, and most of it is free. The backbone is NameBio, a searchable database of hundreds of thousands of reported domain sales pulled from the major auction houses and marketplaces, with data stretching back to 2003. You can filter by keyword, extension, character length, price range, date, and the venue where the sale closed. Alongside it, DNJournal publishes the weekly and year-end charts of high-end brokered sales. Here is an insider detail that tells you exactly what DNJournal is for: a sale generally has to clear roughly 2,000 dollars to even qualify for the weekly chart, so DNJournal is a ceiling-setter for premium and end-user names, not a mirror of the everyday wholesale market.

Treat each source for what it actually measures. Auction venues such as GoDaddy Auctions, Dynadot, Sedo, and NameJet show you wholesale, investor-to-investor prices, and the uncomfortable truth is that the large majority of expired-domain auctions close for under a few hundred dollars. Brokered and retail marketplace sales, the kind that surface on DNJournal or on a marketplace's own sold-listings feed, show you end-user money, which commonly runs three times the wholesale level or more. The same string can be worth 300 dollars at auction and 3,000 to the right company, and the data source tells you which world a given comp came from. Never blend the two without labeling which is which. A comp confirmed through an escrow service such as Escrow.com is a real closed deal; a listing is only a wish.

One more caution: automated appraisals from Estibot, GoDaddy's GoValue, or a marketplace's estimated value are algorithmic guesses, directional at best and frequently wild at the tails. Use them to triage a thousand-name list in an afternoon. Never use them to set a final price, because a robot cannot see the trademark, the buyer, or the story that moved the last comparable name.

• NameBio: the primary free database for searching historical sales by keyword, length, extension, price, date, and venue, going back to 2003
• DNJournal: the verified record of high-value brokered and end-user sales; note the roughly 2,000-dollar floor to make its weekly chart
• Marketplace sold feeds (Sedo, Afternic, Dan/GoDaddy): recent retail transactions inside their own ecosystems
• Auction results (GoDaddy, Dynadot, NameJet, Sedo auctions): wholesale, investor-to-investor levels where most sales close under a few hundred dollars
• Automated appraisals (Estibot, GoValue): directional triage tools only, never a substitute for a real comp
• Your own closed deals and received offers: the single most relevant dataset you will ever own, if you actually record it

Finding a True Comparable, Not Just a Similar Name

A comp is only useful if it is genuinely comparable, and this is where most pricing quietly goes wrong. Two names that share a keyword can live in completely different value tiers. The job is to match on the traits that actually drive price, not on surface resemblance. Search NameBio for your exact keyword first, then widen only along axes that keep the quality class intact. You are hunting for names of the same length, in the same extension, in the same keyword class, and of broadly similar construction quality and commercial intent.

Construction quality is the hard part because it is qualitative. A one-word dictionary .com, a two-word compound, a brandable invented word, and a keyword carrying a hyphen or a number are four different animals even when they all contain the word you care about. The most liquid categories price almost like commodities because so many comps exist: short letter domains such as four-letter .coms, four-number .coms, and single dictionary words trade in tight, well-documented bands. Compare like with like. If you only surface a handful of exact matches, five tight comps beat fifty loose ones every time, because a loose comp silently imports the value of a name that is nothing like yours.

• Same extension: a .com comp does not price a .net, .org, or .io, and the reverse is just as false
• Same length and word count: a 5-letter one-worder is not a comp for an 11-letter compound
• Same keyword class: finance, health, AI, and crypto terms carry very different commercial premiums
• Similar construction: dictionary word, clean compound, or brandable, with no hyphens or numbers unless yours has them too
• Similar end-user demand: a term companies actively brand and budget around beats a term nobody is trying to buy
• Prefer tight liquid categories (short letter/number strings, single words) where dozens of true comps exist

Adjusting a Comp for the Differences That Matter

You will almost never find a perfect twin, so you adjust. Start from the closest comp and move its price up or down for each meaningful difference between that name and yours. Shorter beats longer. A clean, pronounceable, type-in-able word beats an awkward one. A keyword with obvious buyer intent, something a funded company would build a brand around, carries a premium over a term with only casual interest. The extension swing is larger than beginners expect: the same term in .com routinely commands five to ten times what it fetches in .net or .org, which is why cross-extension comps are so dangerous. A hyphen or an embedded number usually cuts value sharply, commonly by half or more, even when every other trait matches.

The honest output of this process is a range, not a single magic number. If your three best comps closed at 1,200, 2,500, and 3,100 dollars, and your name is a touch stronger than the middle one, you are not pricing at 2,500 and you are certainly not pricing at 10,000. You are landing somewhere around 3,000 to 3,500, with a written, defensible reason for each adjustment up or down. Writing the reason forces honesty, because a number you cannot justify out loud in one sentence is a number you simply invented.

Weighting, Verification, and Throwing Out Outliers

Not every data point deserves equal weight. Recent sales reflect the market you are selling into today, so a transaction from the last twelve to twenty-four months counts far more than one from 2011, especially in fast-moving categories where a keyword can triple or collapse in a single hype cycle. Verified sales beat unverified reports, which is why DNJournal-confirmed figures and marketplace-recorded transactions outrank a forum post with no proof. And you actively discard the outliers. One name in your set that sold for ten times the others is almost always a special case: a strategic buyer who already owned the matching trademark, a company that needed that exact string to stop a lawsuit, or a two-party bidding war where ego, not value, set the final click.

Anchor on the median of your comp cluster, not the average, because a single giant outlier drags an average into fantasy while the median stays honest. If you have ten comps clustered between 1,000 and 3,000 dollars and one lone sale at 40,000, the median tells you this is a low-single-thousands market and the 40,000 was lightning. Pricing your name at 40,000 because it once happened to a distant cousin of your name is the single most common way investors strand good inventory for years. Lightning is real, but you do not build a pricing strategy on the assumption that it strikes your name next.

• Weight recent sales (last 1 to 2 years) above old ones that reflect a different, often cheaper, market
• Prefer transactions verified through a marketplace or escrow over anecdotal or unverifiable reported prices
• Discard the top and bottom outliers before you reason about the range at all
• Anchor on the median of the cluster, never the average, so one freak sale cannot distort your whole read
• Record the venue on every comp so you never accidentally mix wholesale auction prices with retail end-user prices

The Gap Between Asking Prices and Sold Prices

The most expensive mistake a beginner makes is confusing asking prices with sold prices. The internet is saturated with domains listed at five figures, and the overwhelming majority of them will never transact at those numbers, or at all. A list price tells you only what one optimistic owner hoped for on the afternoon they set it. A sold price tells you what a real buyer actually paid when money moved. These are different universes, and NameBio is valuable precisely because it shows you closings, not aspirations, which is exactly why it beats scrolling marketplace listings all day.

This gap has a name in the trade: sell-through rate, the share of a listed portfolio that actually sells in a given period. For most investor portfolios it sits in the low single digits per year, often just one to two percent, meaning a 1,000-name portfolio might see roughly ten to twenty sales annually. Asking prices cluster high because listing a name costs nothing but a renewal fee; sold prices cluster far lower because every transaction demands a willing, funded buyer. When you price, you are trying to land inside the sold distribution, not the asking distribution. Studying list prices to set your own number is how you end up with a portfolio that is beautifully, confidently priced and completely, permanently unsold.

Building Your Own Private Comp Sheet

Public databases are the starting line, but your real edge comes from the data no one else has: your own. Keep a private spreadsheet of every sale you close, every offer you receive, and every clean comp you find, logging the name, the price, the date, the venue, and a one-line note on quality and buyer type. Over time this compounds into a pricing engine tuned to your exact niche and your real buyers, not the whole market's noise. Offers matter here as much as sales, because a repeating pattern of 800-dollar offers on names you have mentally valued at 4,000 is the market telling you something important that your ego does not want to hear.

The discipline cuts both ways, and that is the real secret of the trade. A comp sheet stops you overpricing, the loud error that leaves names sitting silent for a decade. It also stops you underselling, the quieter and more common error where you grab the first offer on a name your own records say is worth triple, simply because cash today feels good. Decisions made from a written record are calm, repeatable, and defensible to a future version of you. Decisions made from the gut are where money leaks out of a portfolio, one rushed and unexamined sale at a time.

Frequently asked questions

Is NameBio data accurate, and is the free version enough to price with?

NameBio aggregates reported sales from public auctions and marketplaces, so the recorded transactions are real, but it only captures sales that were disclosed. Many large private and brokered deals never appear, which means the true market can run higher at the top end than the database suggests. The free search is genuinely enough to price most names, since it filters by keyword, length, extension, date, and venue. The paid tiers mainly add deeper filtering, export, and estimate tools, which only start to matter once you are pricing at real volume.

How many comps do I need before I trust a price?

Quality beats quantity every time. Five tight comparables that match your name on extension, length, keyword class, and construction will price it far better than fifty loose ones that drag in unrelated quality tiers. If you can only find two or three true comps, price a deliberately conservative range and say so openly, rather than manufacturing confidence you have not earned. In the most liquid categories, like short letter or number strings, you may find dozens of clean comps, and there your range tightens dramatically. Thin data is a reason for caution, never for optimism.

Why not just price my domain at the highest sale I can find for a similar name?

Because the highest sale is almost always an outlier, typically a buyer who already held the matching trademark, a company forced to acquire that exact string, or two bidders whose egos ran past the actual value. Pricing to the outlier is the classic way names sit unsold for years. Anchor instead on the median of your verified recent comps, then adjust up or down for your name's specific strengths in length, extension, and buyer intent, and treat that one giant sale as exactly what it was: lightning, not the weather you plan around.