In this article
Of all five senses, smell is the only one that skips your brain's gatekeeper entirely — it goes straight to the part that runs emotion and memory. It's also the only sense you don't have a plan for.
You probably have a lighting design — warm white for dinner, not the bright white you'd use for lunch service. You have a music plan with a different tempo at 2pm than on a packed Saturday night. You have a temperature policy, even if it's just the number you leave the thermostat on. For three of your guest's five senses, you've made a deliberate choice.
For the fourth — smell — you probably haven't. There is a smell in your restaurant right now: whatever's coming off the grill's extraction hood, coffee that was just brewed, this morning's laundry load, maybe a scented candle someone bought a year ago that's still burning at the bar. That's not a design. That's whatever's left over when nobody made a decision.
Here's the strange part: smell is the sense with the strongest, best-documented effect on what a guest remembers and how long they stay — stronger than light, stronger than music, in some measurements stronger than taste. Not because scent is magic, but because it's wired differently from your other four senses. This article explains exactly how, with a diagram, further down.
What follows is seven numbers about scent in hospitality — each one independently traced back to its original source, including the study where it simply didn't work. Then you run your own numbers to see whether it's worth the effort for your restaurant.
Why nobody designs this (and everybody has a nose)
Smell is the only sense that reaches your limbic system — the part of the brain running emotion and long-term memory — without a detour. What you see, hear and feel on your skin is first routed through the thalamus: a kind of switchboard that sorts, filters and delays every incoming signal before it reaches the parts of the brain that react to it emotionally. A smell signal skips that step. This isn't a metaphor — it's the wiring, and it's exactly why a smell can drop you back into a kitchen from twenty years ago in a fraction of a second, when a photo of that same kitchen rarely does.
For a restaurant, that means scent isn't just one more dial next to light and music — it's the dial wired closest to the part of the brain that decides how a guest feels about being there. And yet it's the one dial most restaurants never deliberately turn. The ambiance and music planner this site offers covers tempo, volume and lighting for every moment of service — and it deliberately doesn't cover scent, because there wasn't enough to say about it without an article like this one. This is that missing chapter.
There's also a plain, practical reason nobody designs it: a kitchen produces a smell regardless of whether you want it to. What escapes through the extraction hood and what leaks into the dining room is a technical question — governed by standards like the European EN 16282 series for commercial kitchen ventilation — not a design question about what a guest experiences. Those two things get conflated constantly: a well-extracted kitchen isn't a scent policy, it's the absence of a problem.
The 7 numbers
Every one of the seven numbers below traces back to a source you can check yourself — a published study or an EU regulation, not hospitality folklore. Two of them also come with an explicit note on what the number does NOT say, because that matters at least as much as what it does.
1. 15% longer at the table — the first measurement in a working restaurant
Guéguen and Petr set out in 2006 to measure what an ambient scent actually does in a real, operating restaurant rather than a lab: a 22-seat pizzeria in a small town in Brittany, France, over three consecutive Saturday nights between 8pm and 11pm. One night the dining room carried a subtle lavender scent, one carried lemon, one carried none — 88 guests in total across the three conditions (International Journal of Hospitality Management, 25(2), 335–339).
Under lavender, guests stayed at the table 15% longer on average than in the no-scent control condition. Lemon had no measurable effect on how long guests stayed — which is the first thing worth taking from this article: not every pleasant smell works, and "smells nice" is not a guarantee.
2. 20% more spent — same nights, second outcome
The same study also measured what guests spent, and found a similar pattern: under lavender, guests spent 20% more on average than in the no-scent control. Lemon again trailed lavender.
Worth holding onto: this is one pizzeria, one weekend per condition, 88 guests total — not a large-scale study. It is, however, one of the very few scent measurements ever run in an actual working restaurant rather than a simulated retail environment, which makes the number more directly useful to you than most other scent research, almost all of which comes out of retail.
3. 45.11% — the ceiling, not the expectation
The number that gets repeated most often in scent-marketing sales pitches comes from a completely different setting. In 1995, Hirsch measured slot-machine revenue in a Las Vegas casino, comparing a section carrying a specific, pleasant scent ("Odorant No. 1" — its exact composition has never been publicly disclosed) against the same section with no scent. Revenue in the scented section averaged 45.11% higher than the weekends before and after (p < 0.0001), and on Saturdays — when the scent concentration was higher — the lift rose to 53.42% (Psychology & Marketing, 12(7), 585–594).
A second scent tested in the same study ("Odorant No. 2") did nothing at all. That's the pattern running through this entire article: it's never about "a scent" in general — it's about one specific, well-matched scent in one specific context. A windowless, clockless casino floor where people sit in the same spot for hours is also nothing like a restaurant with a 90-minute table turn. Treat this number as the theoretical ceiling of what's been shown possible, not as what your restaurant can realistically expect.
What controlled research actually measured — from nothing to 45%, each in a different context.
The same research question, controlled four separate times, four different outcomes. That's exactly why "scent marketing works" is the wrong promise — context, the scent itself and the execution decide the answer, not the idea on its own.
4. 0 — the number of significant effects in the only real randomized trial
The most honest study in this list is also the least quoted one. In 2016, McGrath, Aronow and Shotwell ran an ambient chocolate scent on and off for 31 days in a real Canadian bookstore-café, randomly assigned day by day — a randomized controlled trial, the strongest research design there is (SpringerPlus, 5(1), 670).
The result: not a single sales figure differed with statistical significance between scented and unscented days (p > 0.10 in every category). Total sales averaged 2.7 Canadian dollars higher on scented days — against a control-day average of 593.5 dollars, which is statistical noise. Book sales actually came in lower (−11.6 dollars), and so did café sales (−8.9 dollars).
This number belongs in every scent-marketing article, and almost never appears in one: a scent that isn't carefully chosen, isn't congruent with what you sell, and isn't placed deliberately, often just does nothing. Put the four numbers above — 15%, 20%, 45%, and zero — next to each other, and they say more together than any one of them alone:
5. No detour — the structural fact behind all four numbers above
Why do four serious studies land so far apart — from nothing to 45%? Partly because scent's wiring is fundamentally different from your other senses, and that difference makes it both more powerful and more unpredictable.
What you see, hear and feel on your skin is first routed through the thalamus — a relay station that weighs, filters and delays every signal before it reaches the parts of the brain that respond to it emotionally. A scent signal travels via the olfactory bulb straight to the amygdala and hippocampus, the seats of emotion and long-term memory — no detour, no filter. That's settled neuroanatomy, not theory.
The result is a double-edged tool. No filter means no delay, no conscious evaluation stepping in first — which is exactly why a scent memory feels so immediate and so emotional. But it also means there's no conscious process correcting for a badly chosen scent on your behalf: a wrong choice of music visibly irritates a guest (and you can hear it and change it), while a wrong scent just doesn't work, with nobody ever telling you why.
What you see, hear and feel passes through a filter first. What you smell doesn't.
Sight, sound and touch are all first routed through the thalamus before they reach the brain's emotion and memory centres. Scent signals travel via the olfactory bulb straight there — settled neuroanatomy, not theory. That explains both why a scent can hit you faster and deeper than music or light, and why there's no built-in correction when the scent is the wrong one.
6. 56 — new fragrance allergens the EU now requires on the label
The moment you deliberately introduce a scent — a diffuser, scented candles, a strongly fragranced cleaning product — you're inside the same regulatory territory as cosmetics. Regulation (EU) 2023/1545, which entered into force on 16 August 2023, adds 56 new fragrance allergens to Annex III of the EU Cosmetics Regulation (EC) 1223/2009 — the list of substances that must appear on the label above a set concentration threshold. That brings the full list to roughly 80 substances, up from the 24–26 it carried before.
This isn't a restaurant-specific rule — it's a product rule for anyone manufacturing or using perfume, scented candles or fragranced cleaning products. But if you choose to run a deliberate scent programme rather than leaving it to the extraction hood, this is exactly the kind of label you should expect from a supplier, and the kind you're entitled to ask for.
7. 31 July 2026 and 31 July 2028 — the two dates that make it real
The regulation carries two transition dates. From 31 July 2026, newly placed-on-market products must carry the expanded labelling — a date that, by the time you're reading this, has already passed. Products already on shelves or in stock before that date get until 31 July 2028 to sell through.
For you as an owner, this changes nothing about what you personally have to do — the obligation sits with the manufacturer, not with you. But it does mean every scented candle, diffuser fluid or fragranced cleaning product you buy from here on should carry a more thorough allergen label than it did a few years ago. For a guest with a scent allergy or sensitivity, that's exactly the kind of information that decides whether their evening ends early.
Do the math: what's scent worth to you — and what does it cost if it doesn't work?
The numbers above range from zero to 45% — which is exactly why "does scent marketing work" is the wrong question. The right one is: at which assumption, applied to your own numbers, is it actually worth doing — and where's the point at which you've just bought expensive candles?
Enter your own covers per week, average spend and the monthly cost of a scenting setup. The assumed-effect slider defaults to a conservative 3% — not the French pizzeria's 20%, and certainly not the casino's 45%, because your restaurant is neither. Change it to whatever number you actually find credible, and see what that means immediately.
The honest scent calculator
No hidden promise — you pick the percentage, the calculator just does the arithmetic.
—
This isn't a promise your restaurant will hit this percentage — the study above with zero significant effects proves the real number could just as easily be zero. It's a calculator to see what's actually at stake at the assumption YOU find credible, not the one that sounds best in a sales pitch.
Compare the payback period above with what you already spend on light, music and temperature — none of those ever gets run through a calculator like this either, even though they do cost something (energy, a sound system, staff time to change the playlist through the day). Scent is the one sense you normally never put a euro sign next to, in either direction.
And if your number comes out at zero, that's not a failure — it's exactly the information most scent-marketing articles never give you, because a zero doesn't sell candles.
Design it, or leave it to your extraction system
You don't have to "do" scent marketing to do something about scent. The first, cheapest step is simple: make sure the extraction above your grill actually works (the EN 16282 standard series exists for exactly this) so that whatever leaks into the dining room is a deliberate choice, not an equipment fault. That isn't a scent programme — it's the absence of a problem, and it's the floor every further choice stands on.
Want to go one step further — a subtle scent at the entrance, congruent with your concept, never masked with a heavy air freshener — run your own numbers above first, with a percentage you actually believe. And if a supplier sells you a scented candle or diffuser fluid, ask for the allergen label that's been mandatory since Regulation (EU) 2023/1545: it's not bureaucracy, it's exactly the information a guest with a scent allergy needs.
You already have a plan for what guests see, hear and feel. These seven numbers aren't here to convince you smell deserves one too — they're here, sourced, so you can decide for yourself whether it's worth it for your restaurant. Including the very real chance the answer is no.