Comparable Sales: The Right Way to Appraise a Domain

Real estate appraisers have used "comps" for a century: to value a house, you look at what similar houses nearby recently sold for. Domain valuation works the same way, and for the same reason — the market's own past behavior is the best predictor of its future behavior. Yet comparable-sales analysis is done badly far more often than it's done well. This article covers how to do it right.

What makes a sale "comparable"?

A comparable is a past domain sale that tells you something about the name you're valuing. The strength of a comparable comes down to three things: relevance, recency, and reliability.

Relevance means the sale shares meaningful features with your name — ideally the same words, in similar positions, on the same extension. A sale of traveldeals.com is highly relevant to traveloffers.com; a sale of traveljohnson.com is not, even though both contain "travel."

Recency matters because the domain market moves. Prices from 2010–2013 reflect a different era; a sale from last quarter reflects today. Recent comps should carry more weight.

Reliability means the sale actually happened at the stated price on a real venue. Rumored prices and unverifiable "I was offered X" stories are noise.

Match on words, not on strings

The most common mistake is matching names by letters instead of by meaning. carsale.com and scarsale.com share five letters in a row but nothing in meaning. Proper comparable analysis first segments both names into words, then matches on those words. This is why word segmentation is step zero of any serious appraisal: it's what lets you find genuinely similar sales instead of coincidental letter overlaps.

Position matters too. The word "app" as a prefix (apploom.com) behaves differently from "app" as a suffix (snapapp.com). Advanced matching accounts for whether shared words sit at the start, middle, or end of the name.

How many comparables do you need?

One comparable is an anecdote. Five start to form a picture. Twenty or more give you a distribution you can actually reason about. The more comps you have, the more you can trust percentiles and the less any single outlier distorts the estimate.

When comps are scarce — say, a name whose rarest word appears in only a handful of sales — your confidence should drop sharply. Thin data is a valuation risk in itself, and it usually signals a thin market, which is a liquidity risk on top.

Weighting: not all comps are equal

Once you've gathered comparables, resist the urge to simply average them. Weight them:

• Give more weight to comps that share more of your name's words.
• Give more weight to recent sales; discount older ones.
• Down-weight comps that share a word but clearly belong to a different category — a "spring" software sale isn't a great comp for a "spring" mattress name.
• Treat extreme outliers with suspicion; investigate before letting them move your number.

A recency-weighted, relevance-weighted view of the market is far more predictive than a naive mean, which a single blockbuster sale can wreck.

The percentile mindset

Instead of asking "what is this name worth?" ask "what's the range this name is likely to sell in, and how fast?" Comparable sales naturally produce a distribution, and that distribution answers both questions:

• The lower quartile approximates a liquidation price — what you'd take to sell within a few months.
• The upper quartile approximates an end-user price — what a patient seller might get from the right buyer.
• The gap between them tells you how much patience is worth, and how uncertain the outcome is.

Communicating a range with a confidence level is far more honest — and far more useful — than a single false-precision number.

Common traps

Cherry-picking the top comp. The most expensive comparable is the one you'll be tempted to quote and the one least likely to repeat. Anchor on the median, not the maximum.

Ignoring category. Shared words can span unrelated industries. Always sanity-check that your comps come from a similar use-case, not just a similar spelling.

Stale data. A comp set that's mostly a decade old will misprice today's market in either direction. Prefer fresh sales and weight accordingly.

Confusing ask prices with sale prices. A name listed at $100k that never sold tells you nothing. Only completed transactions count.

From comps to a decision

Comparable sales give you the evidence; you still have to turn it into a call. Combine the distribution with the name's quality (length, word count, extension, brandability) and its liquidity (how deep the market is for its rarest word). The output should be a range plus a verdict: is the spread between a realistic acquisition cost and a realistic resale price wide enough, and the sell-through likely enough, to make this a buy?

This is precisely the workflow MobiName is built around. It finds the comparable sales for you, positions your name within their distribution, and shows you every transaction it used — so your appraisal rests on evidence you can inspect, not a number you have to take on faith.