The ScentWheel
We measure what a perfume smells like. The ScentWheel reads a fragrance's real notes and plots them on the ten families - the same fingerprint that powers our similarity matching.
In this guide
How the ScentWheel works
A perfume is a blend of notes. The ScentWheel turns that blend into a measured shape you can read at a glance - and compare.
We research every note
We have scored 1,744 perfume notes on 34 fine-grained facets - things like citrus, powdery, smoky, leather and honeyed. Each note gets its own profile, so a real rose and a synthetic rose read differently when they smell different.
34 facets become 10 readable axes
Those 34 facets group into ten families you can actually read: Fresh & Citrus, Floral, Fruity, Green & Aromatic, Sweet & Gourmand, Spicy, Woody & Earthy, Amber - Musk - Leather, Elemental, and Waxy & Fatty. The ten spokes of the wheel are those families. The 34 are still there underneath - flip to "34 facets" to see them.
Breadth - how sure we are
Some notes are precise (a sharp lemon). Others sprawl across many facets (oud, amber). We measure that spread too, and call it breadth. A precise note sits tight on the wheel; a sprawling one casts a wider halo - so the picture also shows how confident the read is.
The shape is the fingerprint
A perfume's wheel is the weighted blend of its notes. Two fragrances with the same shape smell alike - which is exactly how we find dupes and similar scents. We compare the shapes and surface the closest matches.
A family label ("woody") tells you one thing. The ScentWheel tells you how woody, how sweet, how fresh, and how sure we are - all at once, from the real composition. It is measured, not guessed, and the same perfume always makes the same wheel.
See it for yourself
Here is a live ScentWheel. Start with a single note to see its breadth - a focused note sits tight, a broad one opens a wide halo - then switch to a whole perfume to watch its notes blend into one shape. Toggle 10 families for the big picture or 34 facets for the detail.
The drydown journey
A perfume is not one smell - it changes hour by hour as it dries down. The bright opening burns off, a heart settles in, and a base lingers on skin long after. To capture that, we research each note twice: what it smells like, and then when it shows up and how long it lasts.
- When - every note is placed in a phase: the bright top, the heart, or the deep base.
- How long - every note carries an estimated lasting time in hours, produced from its material class rather than measured in a laboratory. Where we hold no value for a note it falls back to the median for its class.
On any perfume page you can scrub the drydown and watch the ScentWheel morph from the opening through the heart to the base, with each note's presence laid out across the timeline above. The ScentWheel is the whole wear, not one snapshot of it.
Most fragrance sites stop at a note pyramid - three fixed lists labelled top, heart and base. That tells you what is in the bottle, but not how it wears. Every note we hold carries an onset and a duration, so the ScentWheel moves through the wear instead of freezing it into three boxes. Those hours are estimates rather than measurements, and we set out how they are produced.
See a drydown in motion on any perfume →
How it relates to other systems
If you have seen a fragrance "wheel" before, it was probably Michael Edwards' Fragrance Wheel - the classic chart that sorts perfumes into families like Floral, Oriental, Woody and Fresh. It is a brilliant map, and our ten families line up with that same intuition.
The ScentWheel takes the idea further. Edwards' wheel is hand-drawn and sorts a whole perfume into one family. Ours is measured from a perfume's real notes, finer-grained, per-fragrance, and it does real work: it powers live similarity. The wider field is moving the same way - towards computational olfaction, where scent is mapped as data (Google's Principal Odor Map is one well-known example). We have built our own corner of that for real bottles you can buy.
We have read the research and we respect the classics, and the recipe is ours: how we weight each note, measure breadth and score a match. We compute profiles across thousands of perfumes, but the quantitative check behind the tuning is small - twenty researched pairs, at a rank correlation of 0.755. We publish that in full, next to the two research groups who did it properly.
Frequently asked questions
What is the ScentWheel?
The ScentWheel is our measured fingerprint for a fragrance. We score a perfume's real notes on 34 facets, group them into 10 readable axes, and plot the result as a wheel. It is the same data we use to find similar perfumes.
Is this AI-generated?
No. The ScentWheel is built by a fixed method from real composition data - the perfume's actual notes and their researched profiles. The same perfume always produces the same wheel. It is measured, not invented.
What does "breadth" mean?
Breadth is how wide or precise a note is. A sharp lemon is narrow; oud or amber spreads across many facets. We measure that spread and show it as a halo, so the wheel also tells you how confident the read is.
How is it different from the Fragrance Wheel?
Michael Edwards' Fragrance Wheel is a hand-drawn family map for whole perfumes. The ScentWheel is measured from each fragrance's real notes, finer-grained, and it powers live similarity matching rather than just labelling.
Can I see it on any perfume?
Yes. Every published perfume on ScentVerdict has its own ScentWheel in the Profile section, with a scrubbable drydown. Browse all perfumes to explore.
How accurate is it?
It is built from researched note data and tuned against a community ground truth of twenty researched pairs, where it reaches a rank correlation of 0.755. That is a small check and we say so. No scent model is perfect, but the ScentWheel gives a consistent read you can compare across fragrances.