In March 2026, Magnum Ice Cream pumped chocolate aroma into a tunnel connecting St Pancras and King's Cross St Pancras stations in London. The tunnel is a major commuter artery — approximately 25 million people pass through King's Cross St Pancras annually, making it one of the busiest interchange complexes in Europe. The installation ran from March 9 through March 22, releasing bursts of chocolate scent at timed intervals, accompanied by a recorded cracking sound meant to simulate the snap of biting into a Magnum ice cream bar.
The campaign was designed, in Magnum's words, to "recreate the multi-sensory pleasure of enjoying a Magnum through sight, sound and scent." The brand called it part of the Magnum Bonbon campaign. JCDecaux installed it. The scheduling and release frequency were calibrated to commuter flow patterns. On paper, this was the commercial exploitation of a perfect transitional space — a liminal corridor where the human nervous system is already in a state of heightened ambient awareness, caught between departures.
What the brief did not account for was the tunnel's existing atmosphere. Transport for London staff at King's Cross St Pancras complained to station management within days. The chocolate scent was drifting beyond the intended installation zone — into TfL break rooms, onto platforms, into escalator corridors. London St Pancras Highspeed, which owns St Pancras station, confirmed that adjustments were made to the timing and frequency of the scent release following complaints. A TfL staff member told The Times: "We've had loads of complaints to us."
The combination of the smell of urine and the chocolate isn't doing it for me. — Advertising professional commuter, speaking to BBC London
The BBC quoted commuters who described the result as "urine mixed with chocolate." On Reddit, one user called it "far too strong and makes me feel a bit nauseous." Another said it smelled "like a chocolate-scented bubble bath — not really appetising at all." One commuter told BBC London the sound "sounded more like something tumbling out of a vending machine."
Magnum maintained, in a statement, that the campaign received "mostly positive feedback from commuters onsite" and would "continue to optimise the campaign based on consumer feedback." The campaign ran its full scheduled duration through March 22. It was not pulled early. It was adjusted.
The tunnel connecting St Pancras and King's Cross is a forced-air environment. The ventilation systems that keep Underground stations habitable also determine how any additional atmospheric element — scent, smoke, aerosol — moves through the space. When you introduce an artificial fragrance into a positive-pressure ventilation system, you are making a prediction about air currents that your brief cannot fully model.
This is the fundamental problem of ambient scent advertising: it is not a controlled laboratory condition. The scent marketing literature has known this for years. A 2013 paper by Doucé and colleagues found that the effectiveness of ambient scent depends critically on the pre-existing olfactory environment — what they called the "ambient odor context." In a space that already smells like diesel, sweat, and damp concrete, a chocolate fragrance does not arrive as chocolate. It arrives as a reaction.
Scent is the only sense that has a direct hotline to the limbic system — the emotional and memory center of the brain — without passing through the thalamus first. This is why it bypasses rational filtering. It also bypasses rational filtering.
The concept of scent congruence — the degree to which an ambient fragrance matches the product or environment it's meant to enhance — has been extensively studied. Aradhna Krishna at the University of Michigan, whose 2012 integrative review in the Journal of Consumer Psychology is considered foundational to the field, established that congruent sensory cues improve brand evaluations, purchase likelihood, and willingness to pay. Incongruent cues can actively harm them.
But congruence is a property of the relationship between the scent and the context, not merely the scent and the product. Chocolate is a perfectly coherent scent for an ice cream brand. It is less coherent when the surrounding atmosphere includes the olfactory signature of 50,000 daily commuters in an enclosed transit tunnel.
The attention economy runs on interruption. Visual advertising — the kind that fills screens and wraps buildings — works by breaking through the schema consumers have developed to ignore routine stimuli. The phenomenon is called banner blindness, and it is well-documented: the human visual system has evolved to filter out predictable patterns, which is what most advertising is.
Scent operates differently. Humans can distinguish approximately 10,000 distinct odors, a range that dwarfs our ability to discriminate colors. Smell triggers 65% recall after one year, compared to 50% visual recall after three months. According to research published in Psychology & Marketing by Spence and colleagues, multi-sensory experiences can increase brand recall by up to 70%. Charles Spence, who runs the Crossmodal Research Laboratory at the University of Oxford, has spent two decades documenting how sensory cues alter perception, judgment, and behavior in ways that visual stimuli cannot.
This is why the commercial scent industry is projected to reach $825 million, growing at 8.2% annually. It's why ScentAir, the global leader in ambient scenting technology, works with brands across retail, hospitality, and transit. It's why Dunkin' Donuts in South Korea ran a campaign that released coffee aroma simultaneously with its brand jingle on buses and saw a 29% increase in sales. It's why, in late 2025, Bath & Body Works became the first brand to run a fragrance ad campaign in New York's Grand Central Terminal — pumping vanilla and pine (branded "Fresh Balsam") into the 42nd Street shuttle platform via diffusers dispersing 20 to 30 pounds of fragrance over the campaign period.
The Bath & Body Works campaign was, by most accounts, a success. Commuters described the experience positively. No complaints were recorded. The difference was not the scent — vanilla and pine are no more inherently pleasant than chocolate. The difference was the environment: Grand Central's 42nd Street shuttle platform has a different ventilation profile, a different existing odor context, and presumably a different demographic of passer-through who is not simultaneously navigating between two major rail hubs while inhaling the accumulated aroma of 50,000 fellow travelers.
The most counterintuitive finding in scent marketing research is what Dipayan Biswas and Allard Szocs documented in a 2019 Journal of Business Research paper: extended exposure to indulgent food scents — the kind used in advertising — paradoxically decreases indulgent food consumption. The mechanism is counteractive-control theory. The scent activates diet goals. The brain, exposed to the idea of indulgence long enough, begins to defend against it.
This is the extended exposure paradox, and it has significant implications for ambient scent advertising in public spaces. A 30-second walk through a chocolate-scented tunnel is a pleasant novelty. A 30-minute shift as a TfL employee in a break room that smells like someone else's dessert advertisement is a labor dispute.
The better ambient advertising works, the more it becomes something people want to escape rather than something they want to stay inside. This is not a failure of execution. It may be a structural feature.
The 2022 replication study published in Frontiers in Communication found that only 20% of influential sensory marketing findings could be successfully replicated in non-WEIRD populations, with larger samples and within-participant designs showing better replication success. The field has a verification problem. Its most celebrated findings are sensitive to context, duration, and population in ways that make universal claims about "what scent does" structurally suspect.
What is more reliably documented is what happens when ambient scent goes wrong. Forced smell exposure is particularly problematic for people with asthma, migraines, or multiple chemical sensitivity (MCS). A 2023 Canadian study in a bookstore-cafe environment found that chocolate scent had no measurable effect on sales of either congruent products (cafe items) or incongruent ones (books) — the context and choice set moderated everything, and the scent itself was statistically indistinguishable from silence. After a couple of months in retail settings, even initially effective scents lose their power as shoppers habituate.
The global outdoor advertising market was worth $40.15 billion in 2025, projected to reach $49.86 billion by 2031. Digital out-of-home (DOOH) advertising — screens, programmatic trading, data-driven placement — accounts for an increasing share of this, growing from roughly $20 billion in 2024 to a projected $39 to $54 billion by 2030. JCDecaux, the company that installed the Magnum tunnel installation, reported 63.5% programmatic revenue growth in 2024.
The commercial scent marketing sector operates as a specialized sub-discipline within this larger apparatus. Prolitec, ScentAir, Air-Scent International, Scentisphere — these companies build the diffusers, design the fragrances, calibrate the release intervals, and sell the promise that your brand can occupy the nose of a consumer in a way that visual advertising cannot. Mastercard has an internal AI-powered sonic station that generates customized sonic DNA tracks. The brand's sonic logo has now been heard at over 590 million terminal interactions. Raja Rajamannar, Mastercard's CMO, has said: "Yet many marketers concentrate on just two: sight and sound. This overlooks some of the most powerful triggers for memory and connection."
The argument is coherent. The neuroscience supports it. Multi-sensory engagement creates stronger neural pathways, generates 50% higher emotional responses, and improves brand recall substantially. The 73% figure — that proportion of global shoppers who believe brands should engage all their senses — appears consistently across consumer research surveys. The commercial logic is sound. The execution is where it becomes interesting.
There is a structural feature of ambient scent advertising that is rarely named directly in the industry literature: it is impossible to opt out of. You can install an ad blocker. You can look away. You can put on headphones. You cannot not smell a tunnel. The scent molecule is already in your respiratory system before your conscious mind registers what has happened.
This is different from visual advertising, which requires gaze direction. It is different from audio advertising, which requires ambient sound levels to reach a threshold. Olfactory exposure is involuntary in a way that other sensory advertising is not. The Reticular Activating System that filters visual and auditory noise does not filter scent before it arrives at the limbic system.
In 2002, Disaronno attempted a similar campaign in the London Underground — wanting to waft an almond-scented fragrance through stations to promote the liqueur. The campaign was cancelled before it ran. The reason, as reported at the time, was that station authorities recognized that hydrogen cyanide gas — the chemical signature of a particular category of terrorist attack — has an almond-like scent. The cancellation was framed as a security measure. It was also, quietly, a recognition that introducing scent into a transit environment creates conditions you cannot fully control or reverse.
The Magnum campaign did not get cancelled. It ran. It was experienced by millions of people who had not consented to it, in a space they were required to pass through, for two weeks in March 2026. Some of them complained. Some of them felt nauseous. Some of them told BBC London it smelled like urine mixed with chocolate. The campaign was adjusted and completed.
Whether this constitutes a failure or a success depends entirely on what metric you are measuring. By the standards of attention capture — which is what the attention economy actually sells — the campaign was probably effective. It generated news coverage, social media discussion, commuter reaction. People talked about it. They are talking about it now. The brand occupied space in a consciousness that was not in the market for ice cream at 8:47 in the morning between St Pancras and King's Cross.
Whether the kind of attention you get from a public that cannot escape your advertisement is the same as the kind of attention you want — whether brand salience achieved through involuntary sensory intrusion has the same commercial value as brand salience achieved through genuine desire — is a question the industry has not yet been forced to answer.
Magnum Ice Cream Company maintains that the campaign was received positively. The tunnel smelled like what it smelled like. The commuters responded accordingly. Somewhere in a lab, a scent marketing machine is being recalibrated for the next enclosed public space, the next fragrance, the next moment of attention that was never requested in the first place.