Within a given consumer market segment, determining a preferred or “winning” sensory stimuli (i.e. flavor, fragrance or cosmetic ingredient) amongst a set of stimuli that, to a layman, have similar characteristics can be a challenge for traditional consumer research. Even large-scale consumer surveys that go into painstaking detail still usually result in a statistical dead heat; likely due to the imprecision such surveys have in representing the subtlety of the consumer’s true emotions. In recent years, this fundamental problem has driven the Consumer and Marketing and Insights team of the Scent and Care division at Symrise to develop alternative research approaches, including a suite of neuroscience methods.
The global market for fragrances and flavors for FMCG stood at US$27.1bn in 2016 and is growing in the mid- to high-single digits. Driven by rapid growth in per capita incomes, Asia is one of the fastest-growing regions for the industry. It is a market in which, from the perspective of the consumer, ingredients with very similar sensory characteristics are pitted against each other with the aim of ascertaining the optimal choice to represent a particular product variant; an ideal scenario in which to apply our suite of neuroscience methods.
In this study, five different fragrances were tested with consumers across Asia for a home care category product: four Symrise candidates and one existing product available in the market. The study examined the sensory aspects in relation to the full product and brand architecture, using EEG, facial coding, eye-tracking and response time testing used to measure non-verbal and subconscious responses of consumers to the interplay of multi-sensory stimuli at various moments of truth. These raw neurometric, biometric and psychometric measures, as well as some traditional survey measures collected for the purposes of validation and comparison were processed by algorithms within our proprietary Gen-isys program into quantitative metrics relating to the affective and cognitive state of respondents.
Apart from addressing the obvious question of which fragrance was preferred by consumers and how explicit measures of liking correlate with implicit measures, we also aimed to understand whether the pattern of relative and absolute liking of the fragrances varied between the different moments of truth. In addition, we wanted to identify if sensorial attributes associations measured with response time tests, correlated well with liking of the fragrance.
Our experience has led us to the understanding that consumer response to sensory stimuli can be skewed by the novelty factor and cultural bias: consumers may initially enjoy a sensory stimulus when they experience it for the first few times, but after repeated exposure, their opinion may change. This is a critical factor, which is traditionally difficult to measure, in determining whether consumers will be back for repeat purchases of the product - quintessential to brand success. In this study, we applied our proprietary “addiction” model, part of our Gen-isys program suite to make predictions about which sensory stimuli will grow on consumers with use.
The figure below shows the liking scores, as determined by our Gen-isys program. Overall, Product E was the best liked, with a performance that was better than both Product A and Product B at the 5% significance level. While the neuroscience-based liking scores did show a modest positive correlation with explicit measures of liking, there was no statistically significant difference between the products when using explicit measures of liking. A similar pattern in neuroscience-based liking emerged when comparing the different moments of truth, although scores for the second moments of truth were generally lower across all products.
In contrast to what might be usually observed in the fine fragrance and beauty care categories, overall liking scores when smelling fragranced products from this category tended to be slightly negative. Nevertheless, our “addiction” model indicated that there was a tendency for all of the products used in this study to be liked more as consumers grow accustomed to them. Again, at the overall level, consumers in China appeared harder to please with respondents there yielding significantly less favorable liking scores across all of the tested products. Product C, which stood out as being of quite a different character from the others, exhibited remarkable variability between countries with respect to liking – this may be a reflection of more seasoned consumers producing lower liking scores, owing to the experience of a wider range of fragrances.
As expected in household products, “fresh” and “clean” were key attributes that correlated well with liking scores. Although many respondents explicitly said the products were unique, response times indicated a lack of conviction. However, respondents did show conviction in their response that fragrances used were suitable for use in this product variant.
This study adds to the existing body of evidence we have built up supporting the use of neuroscience measures as being more sensitive to a consumer’s perception of sensory stimuli versus traditional survey measures. Statistical significance is achieved more readily with neuroscience measures and various analyses indicate improved repeatability. As such, we believe that the addition of neuroscience measures increases precision and provides a more faithful representation of the consumer’s response to sensory stimuli, bolstering confidence to take the actions for the brand.
Jonathan Jacobs (Symrise)