Using maths to mix the perfect perfume
Researchers use online data and network analysis to find the best blend of scents. Mark Bruer reports.
Scientists have used complex mathematical analysis – rather than their noses – to pin down the ingredients for a perfume that smells of success.
Not that they intend getting up the nose of the perfume industry by producing it themselves. Their formula is a suggestion derived from a study of more than 10,000 perfumes aimed at finding out what scents get the best customer responses.
The approach taken by physicists Vaiva Vasiliauskaite and Tim Evans from the Imperial College London, UK, is a novel one for the perfume business.
Perfumes are created by mixing “notes” – individual scents such as jasmine and apple – into combinations called “accords”.
Traditionally, the work of composing perfumes has been a job for “the Nose”, an expert with the knowledge of accords and notes, their volatilities, odour longevities and other attributes important in perfume making.
This expertise is typically acquired over many years of training and trials of many different combinations of naturally occurring oils and chemical molecules. It’s been that way since the earliest perfumes were recorded in ancient Mesopotamia.
But this is the Internet Age, and so Vasiliauskaite and Evans applied a specialised mathematical tool known as complex network analysis to an online database of perfumes, including manufacturers’ descriptions, customer ratings and sales data.
Their first finding was that the popularity of perfumes did not seem to be determined by price or age.
“This motivates us to look at the ingredients, using network methods, to see if these can throw light on what makes a successful perfume,” the researchers write in the journal PLOS ONE.
They found that the 10,599 perfumes in the database used 1047 different notes. While the notes could be identified from the manufacturers’ descriptions, the exact volumes of each note in a perfume remains a commercial secret.
Intriguingly, there was a discrepancy between the most-used notes and those which drew the best ratings from customers.
The most over-represented notes were (in order) musk, jasmine, bergamot, sandalwood and amber. But the notes that appeared to drive up customer ratings most effectively were (in order) anise, orris root, orchid, bamboo and carnation.
Similarly, the most common accords, such as geranium and lavender, were not the ones with the strongest customer response. Lesser-used accords such as jasmine and mint, or musk, vetiver and vanilla, drove much more positive perfume ratings.
“Our results suggest these accords should be more popular than they currently are and that they deserve more attention in the future,” the researchers write.
“Our work provides insights into factors that play a role in the success of perfumes. It also sets up a framework for a statistical analysis of fragrances based on simple properties and customer reviews.
“It could be a beneficial tool for systematic ingredient selection and act as an artificial Nose.”