How we match

A real example

A reviewed alternative is available for Angels' Share. Its comparison shows the available evidence; there are not enough wearer checks to confirm the match yet. Explore the Angels' Share comparison.

What a match means

A match rates how close two fragrances are perceived to smell. It is not a chemical analysis, and it never promises a copy: the words “exact” or “identical” never appear on a result.

Every score is published as a band first — Very close (85% and up), Close (70–84%) or Same family (below 70%) — with the rounded number and the evidence behind it right beside it.

What goes into a score

Four signals, each between 0 and 1. If a signal has no data for a pair it is dropped and the others are re-weighted; nothing is filled in with a guess.

Not used yet: text embeddings and blind tests. Both appear in older plans; they will join only once their data exists, as a new method version.

Opening, heart and dry-down

Many alternatives match the first spray and drift later, so each stage gets its own band: Opening (0–30 min), Heart (30 min–3 h), Dry-down (3 h +).

A stage combines the accords over that time window (50%), the notes listed for that layer (30%) and, for the opening and dry-down, what wearers rated for that stage (20%, once 3 have). A stage with no data shows no band instead of a guess.

Wearers and votes

Signed-in members who have worn both fragrances rate the pair: close match, similar direction or not similar, plus the opening, the dry-down and which lasts longer. One vote per person per pair; a new vote replaces the old one.

An agreement share is shown only from 3 wearers; below that the result says “not enough to confirm yet”, and a percentage is never shown without its count.

Inside the score, experienced contributors’ votes weigh more (up to ten times a newcomer’s), and a vote’s weight halves every 365 days so the score follows the current batch. Daily vote limits and a nightly check for coordinated accounts apply; votes from accounts flagged by that check count neither in the score nor in the tallies these pages show.

Price

Prices are the cheapest listed bottle per ml, from editorial list prices in US dollars collected in September 2026. They are not live and may differ where you shop.

Price only sorts and filters. It never changes how close a pair is rated. There are no affiliate links in these results and no one pays for placement.

How a pair gets published

Anyone signed in can suggest an alternative with a short note on how they compared them. Moderators review suggestions before they appear; suggestions from long-standing expert contributors publish straight away. The site never invents pairs: each one was suggested or added by a person and approved.

An editor can set a pair’s score by hand with a written rationale (for example, when a reformulation made the data stale). Those results say “Editor-set” instead of “Scored on notes & accords”.

If you disagree

Wear both and vote: that is the signal that moves a score. To report a wrong or misleading pair, use “Report relationship” on the pair’s comparison; moderators review every report. A report does not hide a pair on its own — a decision does.

Limits

• Skin, climate and how much you spray change how any fragrance performs.
• Batches and reformulations drift. Votes decay over time so the score leans on recent wear.
• Concentrations differ: search picks the exact edition (EDT, EDP, extrait), and each has its own alternatives.
• Pairs with little data rest on fewer signals — for example, a fragrance without an accord timeline is judged by its notes alone for each stage — and are usually less certain.

Version and changes

match-v1 · 27 September 2026

Scores are computed from accords, listed notes, stage timelines and wearers, recomputed nightly and after each new vote. Before this version, published scores were editorial estimates.