Put the same name into three appraisal tools and you will usually get three numbers, sometimes an order of magnitude apart. Owners tend to read that as noise and pick the figure they like best. Buyers do the same in the other direction. Both are throwing away the most useful information on the screen.
The spread is not random. Each model is built to answer a slightly different question, and once you know which question each one is answering, the disagreement tells you more than any single figure does.
Three questions dressed up as one
Most automated appraisals fall into one of three families. The first looks at the words: length, dictionary status, search volume, how many extensions are already taken. It is really asking how good the name is in the abstract.
The second looks at transactions: what similar names have sold for, adjusted for extension and date. It is asking what the market has paid before. The third looks at demand: inbound enquiries, traffic, operating companies on close variants. It is asking who might pay now.
- Linguistic models flatter short, clean, dictionary names, whether or not anyone needs them.
- Transaction models are only as good as the sales they can see, and most private sales are invisible.
- Demand models move fastest, and are the easiest to fool with a burst of speculative interest.
When the tools agree, the market has already priced the name. When they disagree, someone is about to be surprised.
Which disagreement matters
A high linguistic score with a weak transaction record usually means a name that sounds valuable but has no obvious buyer. It is the most common gap, and the one that keeps names sitting on marketplaces for a decade at prices nobody pays.
The reverse, a modest-sounding name with strong comparable sales, is rarer and more interesting. It often points to a vertical where operating companies are quietly paying for exact matches, and where the next sale will be to an end user rather than another investor.
A demand signal that outruns both of the others deserves the most caution. Sometimes it is real: a new product category, a funded company on a close variant. Often it is a cluster of investors chasing the same idea in the same month.
How we use the spread
PrismIndex scores a name against all three families and keeps the components visible instead of collapsing them into one figure. When we write a valuation for a mandate, the headline number comes with the spread, and the spread comes with a sentence explaining it.
That sentence is usually the part a client remembers. A price can be argued with. An explanation of why the evidence disagrees is much harder to dismiss, and it is what the negotiation is anchored to.
