An essential feature of formulated consumer products is that users generally do not assess their value based on technical specifications, but rather according to functionality and performance attributes, which are often referred to as the quality factors.6 Because quality factors are sometimes qualitative or subjective, performance metrics need to be established. To improve the traditional trial-and-error procedure during this phase, many authors have developed a knowledge basis for the identification of the consumer needs, the inclusion of suitable categories of ingredients, and the desired end-use properties, particularly in the case of cosmetics and personal care products.7–14 In the context of emulsions, Mattei et al.14 proposed such a knowledge basis, including consumer needs, categories of ingredients, and product properties with target values and intervals of acceptance. However, one of the main challenges still to be solved is the formalization of the consumer needs and their translation into measurable properties, namely for product attributes whose assessment is inherently subjective, such as those related to human senses.15 This aspect is particularly relevant for cosmetics, as there is an essential influence of the sensorial attributes interaction with the perception of the user during the assessment of the product.16 To measure this, typically, an objective evaluation (i.e., sensorial profiling) is made by a panel of experts under defined conditions, which makes it a very resource-intensive procedure.17 For this reason, researchers have investigated the characterization of sensory product attributes, mainly translating them to textural and rheological properties.18–20 Nevertheless, these tests are performed by evaluating a single dimension, whereas cosmetic performance evaluation is the result of a combination of sensory, emotional and rational responses during the experience of using the product.21
Since cosmetics or food products are conceptually designed to match consumer expectations, it is crucial to access the implicit emotions people have when interacting with the products and translate them into product specifications.22,23 "Affective Engineering" or "Kansei Engineering" approaches can be used to study how pleasure and efficiency are linked together during the user's assessment of the product.24,25 In such methods, semantic attributes (SA) have been used to describe the subjective responses of consumers to products.26 Traditional techniques, such as PCA (Principal Components Analysis) or multi-factorial analysis, are useful to reduce the semantic space and to quantify the relative importance of each SA or the correlation among them. However, these tools are not suitable to quantify the interactions (i.e., the synergistic or antagonist effects) among groups of SA. For this task, multicriteria techniques, such as the Multiple-Utility Attribute Theory (MAUT) or Analytic Hierarchy Process (AHP), could be used.27 Nevertheless, dealing simultaneously with the set of multidimensional semantic attributes and their interaction is not a trivial task.
Fuzzy measures have been applied to the subjective evaluation of different attributes, such as color or texture.28,29 Recently, a methodology to integrate the user's perception and identify the importance and interaction of consumer attributes, based on fuzzy measures, was proposed by Camargo et al. in the context of massing insoles.30 On the other hand, we had already proposed a systematic method to generate and select alternative formulations, which takes into account available quantitative property models and heuristic formulation rules, with the latter being translated into algebraic restrictions.31,32 In this context, this paper integrates for the first time a quantitative methodology of consumer assessment,30 into a method of computer-aided generation of formulations. The main goal of this approach was to provide a reduced set of feasible alternatives to be prototyped and to accelerate and validate the in-silico design.32 This proposal was partially presented in a short paper,33 being now fully discussed and illustrated, using two examples of skin moisturizers design with the corresponding experimental assessment.