1. Intrinsic quality
The intrinsic model scores extension quality, brandability, commercial intent, clarity, scarcity, build potential, liquidity, buyer depth and naming power. These are properties of the name itself and can be estimated even when no paid data provider is connected.
The score is not a dollar value. A strong domain can still be a bad acquisition if the seller asks too much, and a merely decent domain can be an excellent build when the operator has distribution or monetization advantages.
2. Domain class matters
Premium generics, geo-service names, commercial exact matches, action phrases, descriptive content names and invented brandables do not behave the same way. DomainWorth uses different valuation profiles rather than forcing every name through one generic curve.
3. Market evidence
Live evidence can include RDAP registration history, DNS state, public web presence, cross-extension registration pressure and licensed keyword data. Optional providers are additive. When they are absent, the interface labels the gap instead of filling it with synthetic facts.
4. Valuation ranges
The report separates investor/wholesale value, end-user retail value and a buyer-specific strategic-stretch scenario. It also exposes an acquisition ceiling and suggested selling posture. These are modeled decision ranges, not promises of a transaction.
Worked example: how a score becomes dollars
Here is the full path for domainworth.co, the sample on the homepage. It scores 68/100 and is classed as a commercial exact match. Every figure below comes straight from the model; nothing is hand-tuned.
(68 ÷ 50)⁴ × $205$701× 1.18$828× 1.00$828× 1.10$910$910 × 0.45 to × 1.25$400–$1,100$910 × 3.0 to × 7.0 (class profile)$2,700–$6,500retail high × 1.15 to × 1.75$7,500–$11,500$910 × 0.72$700The score curve is steep on purpose: a 68 and an 84 are not 25% apart in value, they are more than twice apart. Each class has its own retail multipliers, because a geo-service name and an invented brandable sell to different buyers. Rounding is to the nearest $50 under $500, $100 under $5,000 and $500 under $25,000, so the ranges never look more precise than they are.
The math above is the intrinsic model only. In a live appraisal, registration history, cross-extension pressure and any licensed keyword or completed-sale data can move the retail range, and the report labels which kind of evidence moved it.
5. Comparable sales
A small source-linked 2026 sales reference set is included, with reporting dates distinguished from closing dates. This selective set is not a market sample. Matching considers token overlap, name length, TLD, domain class, intrinsic-score similarity and sale recency. At least three close matches (same TLD and class, 0.72+ similarity, at most 8× price spread) are required before comps affect the model. Asking prices may be shown as context but never substitute for a completed transaction. `.cam` has no reported sales in the included reference set; this is missing evidence, not evidence of no market.
6. Confidence is evidence coverage
Confidence describes how much evidence is available to support the range. It is not a sale-probability score. A high-confidence appraisal can still take years to transact, and a low-confidence domain can still sell unexpectedly.
7. Acquisition discipline
Deal Lab adds acquisition price, hold period and renewal cost, then compares the all-in basis with investor and retail scenarios. Domain investing has enough uncertainty already; the model does not add a fake probability-of-sale assumption on top.
8. Portfolio review desk and assumed renewals
Bulk portfolio review groups are generated only from intrinsic scores. A name can appear in several queues: sale-test, build research and high-score hold, for example. Neither the tags nor the dollar estimates prove that a buyer exists or that a domain is unused. The DROP category is always a manual review suggestion, never an instruction to delete or disable auto-renew.
The annual renewal budget lab combines costs supplied in the imported CSV with clearly labeled example rates for missing costs. Those inputs are hypothetical, not real-time renewal quotes, and do not include premium registry renewal charges, taxes, earnings or possible discounts. Merge a registrar CSV by domain for more meaningful figures.
9. Limits
No system can know the future buyer, private negotiations, trademark constraints, seller urgency, buyer-specific strategic need or future market conditions. Use the output to structure decisions and comparisons, not as a guarantee or professional legal advice.