You cannot predict a smell from words, so stop trying. Read for two things only: the base notes, and a stated perfume oil percentage if the house publishes one. Everything else on the page is either unverifiable or actively misleading. Then buy in a way that makes being wrong cheap, which in practice means a discovery set rather than a full bottle.
Why the reading fails: on a measure where 1.00 means everyone gave the same word, English speakers score about 0.95 for naming colours and about 0.20 for naming smells. Descriptions are a lossy channel in both directions, and the words you read before you sniff measurably change what you then smell.
You have a browser tab open. The bottle looks good. The copy says warm, sophisticated, with hints of amber and a soft woody finish.
You still have no idea what it smells like.
Every guide to buying fragrance online gives the same four instructions: read the notes, read the reviews, check the concentration, order a sample. The instructions are fine. What none of them explain is why you need them, and the explanation is more useful than the list, because it tells you which of the four to trust and which to ignore.
Smell does not survive the trip into language and back. Three separate findings say so, and once you know them, buying online stops being a guess and becomes a procedure.
- Words are a poor container for smell. Asked to name 12 odours, English speakers agreed with each other at about 0.20 on a scale where 1.00 means everyone said the same word. Asked to name colours, the same speakers scored about 0.95 (Majid & Burenhult, Cognition 130, 2014).
- You cannot rehearse a smell from a description. Smell imagery is real but unreliable, and the difficulty of summoning it is "a consequence of unstable access to semantic information" (Stevenson & Case, Psychonomic Bulletin & Review 12, 2005).
- The words change what you then smell. Give people the same five odours a week apart with a different label each time and "the perception of an odor could be significantly influenced by the label provided for it" (Herz & von Clef, Perception 30, 2001).
- Even a full ingredient list would not predict the result. There is "no simple correlation" between a material's chemical properties and how it performs, because detection thresholds reorder the hierarchy (Stora, Escher & Morris, Chimia 55(5), 2001).
- So read for the base and the percentage, and sample the rest. Those are the only two published facts that survive the trip to your skin. A discovery set turns an expensive guess into a cheap test.
Does reading about a perfume tell you what it smells like?
Almost nothing, and there is a number for it. In 2014, Asifa Majid and Niclas Burenhult gave the same twelve smells to two groups: ten native English speakers, and ten speakers of Jahai, a language of the Malay Peninsula that has dedicated abstract words for smells. Both groups also named colours. They measured how much the people within each group agreed on the word.
English speakers agreed on colour names almost perfectly and on smell names hardly at all. Jahai speakers agreed on both, at the same level. As the authors put it, "English speakers grappled to find words for odors, Jahai speakers could name odors with the same conciseness and level of agreement as colors."
Two things follow, and they matter for anyone reading a product page.
The first is that the failure sits in the vocabulary rather than the nose. The English speakers could smell perfectly well. They just had no agreed vocabulary, so they fell back on naming the source: this smells like a lemon, that smells like petrol.
The second is that a description is lossy twice over. The copywriter compresses a smell into words with 0.20 agreement, and then you decompress those words with the same 0.20 agreement. Whatever "warm and sophisticated" meant to the person who wrote it, the odds that it means the same thing to you are poor.
And the words do not just fail, they interfere
Rachel Herz and Julia von Clef took five odours with no obvious source, including violet leaf, patchouli, pine oil and menthol, and had people sniff each one twice, a week apart. Nothing about the liquid changed. Only the label did: positive one week, negative the next.
The ratings moved. "Results showed that the perception of an odor could be significantly influenced by the label provided for it", and the authors proposed the cases where a label inverted perception as "the first empirical demonstrations of olfactory illusions" (Perception 30(3), 2001).
Read the marketing copy, then sniff the sample, and you are not running a clean test. You primed yourself. This is the strongest practical argument for ordering the sample before you read too much about it, and for not reading the reviews until after you have made up your own mind.
What a note list can and cannot tell you
A note list is a good ingredient statement and a bad prediction. It is the most useful thing on most product pages, and it still cannot tell you what the fragrance will be like, for a specific and documented reason.
What it does tell you
The base is the part worth reading. It is what remains once the volatile opening has gone, so it is what you will actually be wearing for most of the time the fragrance is on you. Musk, amber, resins, woods and leather-type materials are heavy and stay. Citrus, aldehydes and light green materials leave.
If the base is a list of materials that stay, the fragrance has somewhere to settle. If it is not, it does not. That single reading is worth more than the rest of the pyramid put together.
What it cannot tell you
It cannot tell you the proportions, because nobody publishes them. And even with the proportions, it could not tell you the result. Stora, Escher and Morris measured how fragrance materials deposit and evaporate in consumer products and found "no simple correlation" between a material's physical properties and how it performs, because each material has its own detection threshold, and converting airborne concentration into what a person actually notices "can thus dramatically change the 'physical' hierarchy of the odorants".
In their example, benzyl acetate became undetectable within an hour while phenylethyl alcohol stayed perceptible despite depositing poorly. A note list gives you neither number. It gives you a cast list with no script.
Then there is you. Your skin absorbs part of what you spray rather than only evaporating it, and the balance shifts with how oily and how warm your skin is (Saiyasombati & Kasting, Int J Cosmet Sci, 2003). And the person reading a note list is not the person who wrote it: two individuals differ functionally at over 30% of their odorant receptor alleles (Mainland et al., Nature Neuroscience, 2014). We took that apart in why perfume smells different on everyone, and it is the reason a stranger's five-star review is not evidence about your experience.
Five checks, in order of how much they tell you
Spend your attention on the checks that survive contact with your skin, and skip the rest. Here is the order we would use, and what each one is actually worth.
1. Read the base, and only the base, first
Open the note list and go straight to the last line. If it names musk, amber, patchouli, cedar, sandalwood, leather or a resin, the composition has an anchor. If the last line is citrus or a light green note, you are looking at something designed to be brief. Neither is wrong. Only one of them is what most people mean when they ask whether something lasts. The three tiers, and why the bottom one carries the information, are set out in how to read a fragrance pyramid.
2. Look for a stated perfume oil percentage
A house that prints a number has made a specific, comparable commitment. A house that prints only a category name has given you a price signal. Both of our Eau de Parfums, Golden Hour and Slow Burn, are 50ml at 25% perfume oil, and we publish that figure because it is the one thing on the page you can hold us to. What the words around it mean is your call. What the number means is fixed, although it still does not convert into hours, which is the subject of how to make a fragrance last. We went through why the category name on its own settles nothing in eau de parfum versus extrait de parfum.
3. Treat reviews as weather reports from someone else's country
A review tells you how a fragrance behaved on one person's skin, filtered through one person's receptors, and described with the vocabulary the research above says is unreliable. That is not nothing. It is also not about you. Read them for repeated concrete observations, ignore them for adjectives, and do not read them at all before your own first impression.
4. Check the house has told you where it is made and what is in it
Specificity is cheap to verify and expensive to fake. A brand that names its notes in order, states a concentration, says where the fragrance is made and gives you a way to ask a question is a brand that expects to be checked. We make ours in small batches in India and put the whole note list on every product page for that reason.
5. Buy the smallest quantity that answers the question
This is the check that does the real work, and everything above it is triage to make this step cheaper. You are not trying to decide which fragrance you want from a page. You are trying to decide which two or three are worth putting on your skin. And if there is a specific day they have to survive, test them in the clothes you will wear on it, which is the step nobody takes in what perfume to wear to an Indian wedding.
How to make being wrong cheap
The right question is not "which one will I like", it is "how little does it cost me to find out". The three findings at the top of this piece all point the same way: prediction is unavailable, so buy your way to evidence instead.
Naming figures: Majid & Burenhult, 2014. Prices are ours, current at the date above.
Three days, three bases, one decision
This is exactly what Between Pauses is for. Three 10ml sprays, at ₹1,499: First Impression, After Hours and Bare Skin. Three different bases, so you are testing the variable that matters rather than three versions of the same idea. Wear one a day for three days, on skin, in your own weather. Ask one other person on day two, because you stop noticing your own fragrance within minutes and they do not.
Whichever one you keep reaching for is your answer, and you will have paid ₹1,499 to find it out rather than ₹2,499 to find out you were wrong.
If you already know which direction you want, the full bottles are 50ml Eau de Parfum at 25% perfume oil. Golden Hour opens on peach and pineapple, moves through orange blossom and rose, and settles into musk over patchouli: a daylight scent with somewhere to land. Slow Burn opens with saffron and pink pepper, lets rose soften the heat, then settles into leather, amber, cedarwood and musk. Read those two base lines against check 1 above and you can see what each one is built to do, which is as much as any page can honestly tell you.
Start with the test, not the bottle
Between Pauses is the cheapest honest answer to the question this whole piece is about. Three 10ml sprays, three different bases, on your own skin, for ₹1,499. If one of them turns out to be yours, the full 50ml is waiting.
Frequently asked questions
Is it a bad idea to buy perfume online without smelling it?
It is a bad idea to buy a full bottle that way. Buying a sample or a discovery set online is the sensible version of the same purchase, because it converts a prediction you cannot make into a test you can run. The research is unambiguous that reading will not tell you what something smells like: English speakers agree on smell names at roughly a fifth the level they agree on colour names (Majid & Burenhult, 2014).
Can I trust the note list on a product page?
As an ingredient statement, mostly yes. As a prediction, no. The base tells you whether the composition has materials that stay. It cannot tell you proportions, and proportions would not settle it anyway, because each material has its own detection threshold and those thresholds reorder what you notice (Stora, Escher & Morris, 2001).
Should I read reviews before or after I smell it?
After. Verbal labels significantly shift how an unchanged odour is perceived, sometimes inverting it (Herz & von Clef, 2001). If you read twenty reviews calling something powdery, you have loaded the word before the first spray. Form your impression first, then read, then see whether you agree.
How many samples should I order at once?
Two or three, with clearly different bases. More than that and you run into your own nose: a day of sniffing blurs the comparisons, and you end up rating the order you tried them in. Three is enough to find a direction and few enough to keep the days separate.
Does a higher price mean it will suit me better?
No. Price reflects materials, batch size, packaging and positioning. None of those are about your skin or your receptors. The only page detail with a documented relationship to what you will experience is the base, and the only number you can compare across brands is a stated oil percentage.
What about buying for Indian weather specifically?
That is a separate variable and a real one. Humid air keeps skin damp, which changes how a fragrance sits and travels, and Indian humidity varies enormously by city and season. Our guide to choosing a perfume for Indian weather works through it with the climate data, and it pairs with this piece: that one tells you what to look for, this one tells you how much to trust what you read.
The short version
Reading cannot tell you what a perfume smells like, and the reading can make your first impression worse. English speakers agree on smell words at about a fifth the level they agree on colour words. Descriptions lose the smell going in and again coming out, and a label you have already read will tilt the sniff that follows it.
So use the page for the two things it does carry honestly. Read the base to see whether the composition has an anchor. Look for a stated oil percentage, because a number is the only claim you can compare between brands. Then stop reading and start testing.
Between Pauses exists for that step: three 10ml sprays, three bases, ₹1,499. If you already know your direction, Golden Hour for daylight and Slow Burn for after dark, both 50ml at 25% perfume oil. Either way, the bottle that wins is the one someone asks you about.
Remembered, not recognised.
The naming figures come from a published study whose sample size we state on the chart, and the values are read from that paper's own figure and rounded, which the chart says on its face. The labelling study is quoted directly from its abstract. Where a classic source could not be read in full, we left it out rather than repeat a number secondhand. The matrix in this piece is our own assessment, labelled as such, and the reasoning behind each mark is in the section above it.
We do not publish wear-time, projection or spray-count claims, and we do not display ratings we cannot stand behind. Editorial: written by the Bare Affair house team, every figure checked against a named source, last reviewed 18 September 2026. If you spot something wrong in it, tell us and we will correct it.
More about the house, and how we make our fragrances: About Bare Affair · Contact us · WhatsApp concierge: +91 96666 17181
Sources
- Majid, A. & Burenhult, N. (2014). Odors are expressible in language, as long as you speak the right language. Cognition, 130, 266–270. doi.org/10.1016/j.cognition.2013.11.004
- Herz, R. S. & von Clef, J. (2001). The influence of verbal labeling on the perception of odors: evidence for olfactory illusions? Perception, 30(3), 381–391. doi.org/10.1068/p3179
- Stevenson, R. J. & Case, T. I. (2005). Olfactory imagery: a review. Psychonomic Bulletin & Review, 12(2), 244–264. doi.org/10.3758/BF03196369
- Stora, T., Escher, S. & Morris, A. (2001). The physicochemical basis of perfume performance in consumer products. CHIMIA, 55(5), 406–412. doi.org/10.2533/chimia.2001.406
- Mainland, J. D., Keller, A., Li, Y. R., et al. (2014). The missense of smell: functional variability in the human odorant receptor repertoire. Nature Neuroscience, 17, 114–120. doi.org/10.1038/nn.3598
- Saiyasombati, P. & Kasting, G. B. (2003). Two-stage kinetic analysis of fragrance evaporation and absorption from skin. International Journal of Cosmetic Science, 25(5), 235–243. doi.org/10.1046/j.1467-2494.2003.00193.x