How We Score Fragrance Clones

How closely the clone smells to the original. This accounts for note-by-note fidelity, overall accord match, and the opening, heart, and dry-down stages. A score of 9–10 means most people in a blind test would struggle to distinguish the two.

How long the clone lasts on skin compared to the original. Many clones use different aroma-chemical concentrations, which can mean dramatically better or worse staying power. Scores factor in both sillage (projection) and skin longevity.

Price-to-performance ratio. A clone that smells 85% as good for 5% of the cost scores extremely high here. Value Score rewards accessible pricing — a $15 clone doesn't need to beat a $350 original to earn a 9.

The single headline number: 40% Similarity + 30% Longevity + 30% Value. This powers the rankings on Best-Of listicles and the default sort on fragrance pages.

When you trigger a comparison, Claude receives a structured prompt containing the full note pyramids of both the original and clone — top notes, heart notes, and base notes — along with their concentrations and the clone house's known quality profile.

Claude analyses ingredient overlap, accord structure, and likely dry-down trajectory to generate a verdict and assign all four scores. The output is structured JSON: verdict text, plus four numeric scores rounded to one decimal place.

To keep costs low and responses fast, results are cached in our database. Once a comparison has been run once, all subsequent visitors see the cached result instantly — no API call needed.

Our AI scoring is informed by published olfactory research including work from Osmo.ai and MIT Smellnet.