Why automated domain appraisals mislead — and what to use instead

Automated appraisal is genuinely useful for one job: separating names with some value from names with none, quickly, across a large list. Used for anything beyond that it produces confident numbers with no reasoning attached, which is worse than no number at all because it anchors the conversation.

What the models actually see

Character count. Extension. Whether the string appears in a dictionary or a keyword list, usually an English one. Rough registration and expiry dates. Sometimes a pattern match against a sales database.

That is a reasonable set of inputs for an English two-word .com in a well-traded category, and those are the names the models were fitted on. Move outside that and the inputs stop describing the asset.

Where they break on Vietnamese

A Vietnamese term stripped of its tone marks does not appear in any English dictionary, so the model sees a random string and scores it near the floor. Meanwhile the same string is an ordinary category word to sixty million people.

The reverse error happens too. A string that looks like a plausible English brandable may be meaningless or unfortunate in Vietnamese, and the model has no way to know. Both errors are systematic rather than random, which means you cannot correct for them with a fudge factor.

The narrower failure everyone hits

Even within English, automated appraisal cannot see commercial intent. It does not know that a term is searched by buyers rather than browsers, that a sector is consolidating, or that one company has spent two years trying to acquire the name. Those are the facts that actually move a price.

It also cannot see the thing that determines whether a sale happens at all: whether a specific buyer exists. A name with one obvious owner and no substitutes is worth a great deal to that owner and very little to anyone else. No model has a column for that.

The method that works instead

Read the name. Establish what it means, in the language it is in, confirmed by someone who speaks that language. Everything downstream depends on this and nothing else can substitute for it.

Name the buyer. Describe the specific kind of business that would want it and what they would use it for. If you cannot write two sentences describing that buyer, you do not have a price problem, you have an inventory problem.

Find the alternatives. What else could that buyer use, and what would it cost them? A name with three close substitutes is priced by the substitutes. A name with none is priced by what replacing the brand would cost.

Cross-check with four estimates. Replacement cost, adjacent-market comparables, search demand, and revenue anchoring. Discard the highest and lowest; if the middle two agree, that is your number. If they disagree by an order of magnitude, stop and work out why before quoting anything.

When the automated number is worth having

When you are screening a list of two hundred names and want to know which forty deserve an hour of attention. When you want a rough floor for a bulk portfolio. When a counterparty has quoted one at you and you want to know what they are anchoring on.

Never as the price on a listing, and never as the justification in a negotiation. “A tool said so” is not an argument, and any buyer who has been in the market longer than a month knows it.