How we price five thousand domains
Listen to this post
AI narration.
Nobody can hand-price five thousand names honestly. Gut feeling does not scale, and it flatters the names you are fond of. So our prices come from a pipeline, and this post is what is in it.
Several independent signals. Each name is appraised by multiple outside providers, independently of each other. We pay for these numbers; we do not produce them.
A blind referee. Where providers disagree hard, a language model adjudicates — blind, meaning it sees the name but not which provider said what, so it cannot learn to flatter one source. Its verdicts were checked against our own recorded sales before it was allowed to vote.
Outliers are clipped. One provider having a wild day should not move a price much. Extreme signals are pulled toward the field before blending, symmetrically — highs and lows alike.
The blend becomes the price. The blended value produces the Buy Now figure, and every derived amount — monthly lease payments, minimum offers — is computed with half-up rounding in integer arithmetic. No floating-point dust, no banker's rounding surprises.
Every price has a history. Pricing logic is versioned, and each published price records which version produced it. When we change the formula, the change ships like code: reviewed, tested against the previous version, and auditable afterwards.
None of this makes a price "correct" — only a sale does that. But it does mean the number on the page is the same number we would compute for a stranger's name, and that we can tell you exactly how it came to be.
Comments
Comments are reviewed before publication. The Author badge marks replies from the author.
Your test session has expired. Stop testing to use comments and sign-in normally, or open a fresh test link to keep testing.
Stopping testing reloads this page in normal mode.
Read approved comments on the discussion page.
Open the discussion page