1. Introduction
Mazamorra morada is a traditional and widely consumed dessert in Peruvian cuisine. This dessert is commonly prepared at home, at local restaurants, or sold as a street food, and it is consumed fresh all over Peru. Mazamorra morada stands out because of its intense dark purple color, which is provided by its main ingredient, the purple corn (Zea mays L.), a variety of Peruvian Andean maize, (Salvador-Reyes & Clerici, 2020). Besides of the corn, the other ingredients used for preparing mazamorra morada include potato starch (which is used to give consistency to this dessert), chopped fresh fruits (pineapple and peach), dried fruits (raisins and prunes), spices (cinnamon and clove), and sugar (Saldaña et al., 2018). Mazamorra morada is highly popular in Peruvian eating habits and is almost exclusively sold as a street food.
Certainly, hearing “the consumer’s voice” in the initial stages of a new product development is key to guaranteeing its success in the market (van Kleef et al., 2005). However, this step is frequently ignored or misunderstood in the development of certain products. Currently, consumers are more and more demanding in their choices and preferences, looking for products that satisfy aesthetic attributes beyond sensory characteristics (Saldaña et al., 2020). According to Köster (2009), consumer food choices are strongly influenced by six main factors: (i) intrinsic product characteristics, (ii) biological and physiological factors, (iii) psychological factors, (iv) situational factors, (v) socio-cultural factors, and (vi) extrinsic product characteristics. This study highlights the extrinsic characteristics of the product, as they must be carefully chosen when designing a package vying for success in the market. The extrinsic factors, known also as non-sensory factors, are “related to the product but do not form part of the physical product itself” (Ampuero & Vila, 2006), being the first factors to be put in contact with consumers and having a proven influence on their food choices (Ares et al., 2009). To date, only studies on sensory properties (intrinsic properties) associated with mazamorra morada have been carried out (Saldaña et al., 2018) but no study on the non-sensory properties (extrinsic properties) have been conducted. Extrinsic factors mostly comprise label, price, origin, convenience, nutritional information, and health statements, among others (Gil-Pérez et al., 2019; Sáenz-Navajas et al., 2013). Moreover, the most important non-sensory factors for food choice are convenience, price, production technology, personal health, branding, social and political issues, and contextual influences (Jaeger, 2006).
Usually, the nutritional composition of food products is informed to consumers through nutritional labeling on the packaging (Ares et al., 2018). Recently, the front-of-package (FOP) labeling typology has stood out because it simplifies consumers' understanding of the nutritional content of food products by using logos, symbols, and icons (Dean et al., 2014). There are different types of FOP labeling (Aschemann-Witzel et al., 2013), among which the warning type presents significant advantages over the others in terms of reaction time of consumers, thus facilitating the informed purchasing decision (Arrúa et al., 2017a). Another important non-sensory factor is "quality", which has a broad and multidimensional definition with both objective and subjective components in each situation and time frame (Lawless, 1995). The objective component refers to physical characteristics present in the product and the subjective component is associated with how the consumer perceives quality (Grunert, 2007). Interestingly, the concept of quality becomes more important when there are physical modifications on the product resulting in a high-quality product for consumers. According to (Shepherd and Raats, 2010), the quality expectation is formed by quality cues, such as some texts on the label referring to quality or packaging materials that reflect quality. The packaging material has been scientifically proven to have a strong influence on consumers’ perception about the product, which is expressed by their intention to purchase it (Rebollar et al., 2017). Similarly, another non-sensory factor, the convenience that a product offers to the consumer, in terms of practicality, can significantly affect the expected acceptance (EA) and the purchase intention (PI) of a product (Saldaña et al., 2020). Therefore, products with low convenience are not well-valued by consumers who do not want to spend time/effort in using or consuming that good or service (Buckley et al., 2007). Conversely, if a product is perceived as practical, the PI and re-purchase will be high (Hyldelund et al., 2020). Finally, the naturalness of foods is a broad construct that includes the following aspects: "natural content" (Steptoe et al., 1995), "without additives" (Roininen et al., 1999), and “natural foods” (Roininen et al., 1999). Despite the different nuances of naturalness, it is known to be a positive attribute for consumers. However, its importance varies depending on the culture, country and history and a given product or service (Román et al., 2017). In the case of mazamorra morada, which is either a street food or a homemade dessert, it is mainly consumed warm once prepared, so that the perception of “natural” is implicit. The development of a ready-to-eat mazamorra morada, certainly, would negate the distinctiveness of a "homemade" dessert, making the product less natural for certain consumers. This, undoubtedly, would negatively impact the intention to purchase and eat it. For this reason, incorporating naturalness as an extrinsic factor (through a picture of purple corn) to this study is very important.
The expectations generated by the different non-sensory factors need a reliable design and analysis that allows us to measure the impact of these factors and their levels of interest. One of the most used and established techniques to measure the impact of non-sensory factors on sensory and consumer science is the Conjoint Analysis (CA) (Claret et al., 2012). This technique allows researchers to estimate the structure of consumers’ responses by using a set of profiles of a target product, which is obtained by pre-determined combinations using a specific experimental design (Green & Srinivasan, 1990). Understanding the structure of consumers’ responses will help to understand the factors that modify the EA, PI, and holistic perception (HP). The HP is based on the evaluation of global similarities and differences among products (Varela et al., 2017) rather than common analytical descriptions and hard‐to‐verbalize characteristics (Valentin et al., 2018). Once the relative importance and the utility of levels of each factor have been estimated, it will be possible to create packages and labels that capture the interest of consumers and promote an informed purchase decision of a product (Ares and Deliza, 2010). Several studies have investigated consumer behavior towards products with several modifications to their label (Rojas-Rivas et al., 2020), brand (Varela et al., 2010), shape (Gislason et al., 2020), color (Baptista et al., 2021), figures (Ares et al., 2016), nutritional information (Gębski et al., 2019), health claims (Gislason et al., 2020), ecological aspects (Lim et al., 2018), price (Calegari et al., 2018), and processing technology (Martins et al., 2020).
To our best knowledge, there are no previous studies addressing the impact of extrinsic factors on consumer behavior towards ready-to-eat mazamorra morada. In this context, the aim of this study was to estimate the salient extrinsic factors (packaging material, nutritional information, quality, convenience, and naturalness) in terms of EA, PI, and HP for mazamorra morada, based on consumers’ perception.
2. Materials and methods
2.1 Stimuli
Stimuli were designed by the combination of factors and levels previously selected, based on a pre-test with regular consumers of mazamorra morada. For that, ninety-one target consumers (frequency of mazamorra morada consumption = once a month or more, Mean age = 24 ± 4) answered, through Google forms, the following openended question: “Please, write down all components (words, descriptors or associations) that come to mind when you think of the packaging and label of mazamorra morada” (Piqueras-Fiszman et al., 2013). After the catego rization of components, 16 semantic categories were ob tained (Figure 1), with the most frequent being: nutritional information, packaging material, expiration date, label, brand, nutritional composition, and convenience.
Of these most important factors, the expiration date and label were kept as constant factors in all the stimuli, because they appear in all ready-to-eat products in the Peruvian market. Thus, the five factors to study were: packaging material, nutritional information, quality to express the brand (until the execution of the study, there was no brand that produces ready-to-eat mazamorra morada, avoiding the selection of different brand levels), convenience and naturalness. These factors were used to create 12 stimuli by means a fractional factorial design considering two levels per factor as shown in Table 1. Concerning “packaging material”, they ranged from food grade paper (FGP) to glass, since these were the levels most frequently cited by consumers in the open-ended question. Regarding “nutritional information”, the levels ranged from the classic “Guideline Daily Amount (GDA)” system to the food warning system, currently used in Chile (Ministerio de Salud de Santiago de Chile, 2015), Uruguay (Ministerio de Salud Pública de Montevideo, 2018), and Peru (Ministerio de Salud del Perú, 2018). The warning system consists of placing a “High in ‘X’” or “Excess ‘X’”, where X would refer to the nutrients of concern when they are high/excessive. This system was recommended by The Pan-American Health Organization (PAHO), since it presents visible, clear and concise information to consumers (Pan American Health Organization, 2016). Finally, the low level for quality was the absence of any cue while the high level was designated ‘premium’. Regarding convenience and naturalness, the low and high level considered the addition of napkins and a spoon, and no image and the addition of an image, respectively. Incorporating the study of naturalness is important since consumers have a strong preference for natural foods, being a "decisive buying incentive" because natural products are perceived as healthier (Román et al., 2017).
Factor | Low level | High level |
Packaging material | FGP | Glass |
Nutritional labeling | GDA | Warning |
Quality cue | None | Premium |
Convenience | None | Spoon and napkin |
Naturalness | None | Image |
GDA: Guideline Daily Amount. FGP: Food Grade Paper.
To estimate the relative importance and utility of the five two-level factors without increasing exponentially the number of stimuli, an orthogonal fractional factorial design was used (Table 2), assuming negligible interactions of factor effects (Claret et al., 2012). All combinations were used to design the graphical stimuli using Adobe Photoshop software (Figure 2).
2.2 Consumers
The questionnaire was sent via email to students, professors and officials from the Agroindustrial Engineering School of the Universidad Nacional de Trujillo (UNT), inviting them to participate in the study of “mazamorra morada” through a link with the questionnaire implemented in Compusense Cloud software (Compusense Inc., Guelph, Ontario, Canada). Before answering the questionnaire, the participants were allowed to accept or decline their participation in the study through an informed consent form. This study is in accordance with the declaration of Helsinki and was approved by the UNT ethical committee.
The within-subjects online questionnaire was answered by 300 Peruvian consumers. However, only 97 responses were valid (questionnaires not completed or exceeding 10 minutes’ time that reflect neglect). Out of the 97 respondents considered in this study, 49 were male and 48 were female, ranging from 18 to 57 years old (mean age = 26). Regarding consumption frequency, 56% claimed to consume mazamorra morada once per month, 27% stated to consume it once per week, 14% twice per week, and 3% at other frequencies.
Stimuli | Packaging material | Nutrition labeling | Quality cue | Convenience | Naturalness |
1 | FGP | GDA | None | None | None |
2 | FGP | GDA | Premium | None | None |
3 | Glass | Warning | Premium | None | None |
4 | Glass | GDA | None | Spoon and napkin | None |
5 | Glass | Warning | None | Spoon and napkin | None |
6 | FGP | Warning | Premium | Spoon and napkin | None |
7 | FGP | Warning | None | None | Image |
8 | Glass | Warning | None | None | Image |
9 | Glass | GDA | Premium | None | Image |
10 | FGP | GDA | None | Spoon and napkin | Image |
11 | Glass | GDA | Premium | Spoon and napkin | Image |
12 | FGP | Warning | Premium | Spoon and napkin | Image |
GDA: Guideline Daily Amount.
2.3 Procedure
Firstly, consumers indicated their sociodemographic profile and consumption frequency of mazamorra morada. Then, visual stimuli were coded with 3 random numbers and presented to consumers in a sequential monadic format, following the Latin square design. Subsequently, consumers were asked to indicate their EA for each stimulus by means of a linear hedonic scale of 10 cm ranging from “dislike extremely” on the left to “like extremely” on the right. Then, consumers were asked to indicate their PI by using a categorical scale of 5 points (from “I would definitively not buy it” to “I would definitively buy it”).
Finally, the Napping test was applied (Figure 3) to know the HP that stimuli generate in consumers’ mind (Lê et al., 2015). To do so, all stimuli were shown to consumers on the right side of the screen and respondents were asked to allocate the stimuli in the box showed on the left side of Figure 3, according to their similarities and differences (Saldaña et al., 2019). This was done considering that two stimuli which are close together are similar, whereas if they are far apart, they are different (Varela & Ares, 2012). Consumers were encouraged to use the whole surface of the box and, when necessary, the consumer could zoom in on stimuli by clicking the symbol twice for a better visualization. Finally, consumers were asked to indicate a few words to describe each stimulus allowing us to understand the reason of the similarities and differences between stimuli.
2.4 Data analysis
EA and PI of the stimuli were evaluated through an analysis of variance (ANOVA), considering stimulus, consumers, and presentation order as factors. A Tukey's test was used for pairwise comparison of stimuli. Both analyses were carried out at 5% significance level.
For CA, the relative importance of each factor and the utility associated to each level were estimated. An individual additive model was used to determine the utility of each level. This model assumes that the assessments of expected liking/purchase intention are formed by the sum of the contributions of the factor levels as recommended by (Lima Filho et al., 2015). The structure of this model for m j factors and n levels is shown in equation (1):
Where, Y was EA or PI for a stimulus, V ij is the utility corresponding to j th level for i th factor (i = 1, 2, 3…n and j = 1, 2, 3.m), X ij is the variable that indicated the presence of j th level for i th factor in the stimulus, and Ɛ ij is the random error of the model. Once the utility of each level was determined (Vij), the relative importance of each factor was determined according to the equation (2). Where, I n is the importance of i th factor. The sum of the importance of all factors accounts for 100% (Raz et al., 2008).
In addition, a Multiple Factor Analysis (MFA) was performed on the Napping data (Pagès & Husson, 2013). Initially, the coordinates of each stimulus were obtained from the perceptual map of each consumer, using the lower left side of each perceptual map as the origin of the coordinates. Consumers’ coordinates were considered as active variables and stimuli descriptions as supplementary variables. Confidence ellipses were built by parametric bootstrapping at 95% significance to evaluate the stability of samples in the perceptual map by using the script provided by (Dehlholm et al., 2012).
All analyses were performed using R software (R Core Development Team, 2019). The ANOVA was performed using the lmerTest package (Kuznetsova et al., 2017). The relative importance and utilities were performed using the Conjoint package (Bak et al., 2018) and the MFA of Napping was performed using FactoMineR package (Lê et al., 2008).
3. Results and discussion
3.1 Expected acceptance and purchase intention
The results of open-ended questions indicate the most relevant factors and levels for CA based on the frequency of elicitation of semantic categories (see Table 1 and Table 2). This selection was validated by the discrimination of stimuli by individuals for EA and PI (see Figure 4 and Figure 5, which show the significant difference between stimuli for EA and PI). On the one hand, the stimulus 10 presented the highest EA, being significantly higher than the stimuli 1, 3, 4, 7 and 8. On the other hand, Stimuli 3 and 8 presented the lowest EA.
Regarding PI results, the pattern found was similar to that of EA. Stimuli 6 and 10 presented the highest PI, which was significantly alike for both, and they were only higher than stimuli 1, 3, 8 and 9. Stimulus 8 presented the lowest PI. Therefore, both EA and PI were affected in a similar way by the different levels of factors.
EA and PI are important variables to be measured when designing a product because they provide quantitative responses for consumers’ perception of a product and give a projection of the possible success of a product in the market (Costell et al., 2010). Based on the images of the stimuli placed in the upper part of Figures 4 and 5, they allow us to affirm that convenience, packaging material, and quality cues significantly impacted EA and PI.
The presence of a “napkin and spoon”, as well as a package made with “FGP” increased EA and PI at the same level. Convenience has proven to be a determining factor in food selection, increasing the intention of purchase. This increase is explained by the short time invested to consume or use a product (Bove et al., 2003). However, any improvement in convenience will impact food choices in different ways. For example, Saldaña et al. (2020) reported that convenience did not significantly impact the EA and PI of smoked bacon, probably because this type of product was not ready-to-eat after purchasing and involved an additional preparation process. For the case of mazamorra morada, we hypothesized that convenience had a greater impact since it was studied in a ready-to-eat product. However, convenience can also be explained by the current consumption of this dessert in Peru, which is frequently consumed as street food (Valentin & Gomez-Corona, 2018). In other words, both street food and ready-to-eat products can be perceived as convenient by the consumer. The convenience of using napkins and a spoon could be enhanced by using a light and resistant material, such as FGP.
3.2 Conjoint analysis
To fully understand the impact of factors and levels of stimuli on EA and PI, the relative importance and utility of each factor and their respective levels were determined. It was observed that the relative importance of EA and PI was similar (Figure 6A), meaning that consumers rated both parameters similarly. The EA and PI of consumers showed to have the same behavior pattern. Although PI and EA do not represent the same construct (as PI is more related to spending money to buy a product), the similarity in consumer’s behavior towards these parameters may be associated with the fact that the consumer pretends to spend money to buy a product that manufacturers think the consumer will like (Saldaña et al., 2020).
For Peruvian people, packaging material was the most important factor, followed by convenience, naturalness, quality, and nutritional labeling (Figure 6A). The two most salient factors were “packaging material” and “convenience”, which presented a relative importance higher than 20%. It is known that the packaging material can generate expectations on consumers, increasing their intention to buy a product (Rebollar et al., 2017). This tendency was confirmed for mazamorra morada and, looking to the utility values (Figure 6B), it is possible to indicate that FGP material was the level responsible for the increase of EA and PI, while the glass level was responsible for the decrease of EA and PI. This behavior can be explained by the previous experience of consumers regarding the superior practicality of FGP package when comparing to the glass one. For convenience factor (the second most salient factor), the incorporation of a spoon and napkin significantly increased utility values, showing their importance for consumers when eating mazamorra morada. This can be explained by previous experiences of consumers when eating mazamorra morada, as this dessert has a consistent texture, which makes it difficult to eat without the help of a spoon and a napkin. In addition, mazamorra morada is ideally eaten warm. Thus, the use of a spoon and napkin is almost essential in terms of practicality and for avoiding burn accidents.
The remaining factors (naturalness, quality, and nutritional labeling) presented a moderate relative importance close to 15%. According to the levels of factors, the use of an image of purple corn increased the utility of naturalness and the incorporation of the word "premium" positively influenced the utility of quality. Finally, the nutritional labeling style shows that the warning system decreased PI, showing once again the effectiveness of this system to warn consumers about nutritional danger; for example, mazamorra morada is "high in sugar". This finding is in line with other studies on a wide variety of foods, for which the warning system was used in their nutritional label (Antúnez et al., 2020; Arrúa et al., 2017b; Deliza et al., 2020; Khandpur et al., 2018). It is important to highlight that this study is the first of its kind considering the perception of Peruvian population regarding the warning system used in food, which should be explored in depth with further studies to design public policies based on empirical evidence on the use of this system.
Therefore, these results bring out the levels with a positive utility of factors that consumers considered relevant for a ready-to-eat mazamorra morada. These were a FGP package (packaging material factor) with a spoon and napkin (convenience factor), as they presented the highest EA and PI, and these were represented by the stimuli 6 and 10, validating the results analyzed by ANOVA (using stimuli as factors).
3.3 Napping
To explore another aspect of consumers’ perception, the Napping test was used to access the holistic perception of consumers, which is much more intuitive and, therefore, less analytical than EA and PI (Varela et al., 2017). Holistic perception is worth studying since it shows less analytical nuances, typical of the rapid decisions made by most consumers in a supermarket (Saldaña et al., 2019). Based on confidence ellipses, the stimuli were clustered in three main groups (Figure 7). The first cluster was comprised of the stimuli 1, 2 and 7, which were perceived as natural and healthy, and presented an intermediate EA and PI. This means that, despite consumers stating that they worry about their health, there are other factors more important that may determine their purchase decision. This may be associated with the fact that most of the respondents were young adults, who are generally less concerned about their health (Ares et al., 2010). The second cluster was comprised of stimuli 6, 10 and 12, which were perceived as convenient and high quality, also presenting the highest EA and PI.
These stimuli were designed by using FGP as packaging material, and a "napkin and spoon" as convenience level. In view of this, for the future processing of mazamorra morada, one of these stimuli should be considered as a possible prototype to be launched to the market. The third cluster was composed of the remaining stimuli, which were in the positive part of the first dimension of MFA and presented the lowest EA and PI. This last cluster represents designs of glass containers.
4. Conclusions
Non-sensory factors impacted significantly on EA, PI, and HP. Convenience and packaging material were the most salient factors, presenting the highest relative importance. Based on the utility of levels of these two factors, the use of a FGP as packaging material, and the inclusion of a spoon and napkin increased EA and PI of the mazamorra morada package. These findings were confirmed by stimulus 6 (Packaging material = FGP, Nutrition labeling = Warning, Quality cue = Premium, Convenience = Spoon and napkin, Naturalness = None) and stimulus 10 (Packaging material = FGP, Nutrition labeling = GDA, Quality cue = None, Convenience = Spoon and napkin, Naturalness = Image), which presented these levels and are the most well-rated by consumers. In addition, according to the Napping results, the stimuli were clustered in three main groups by consumers, using, as the main criteria, convenience and packaging material. Thus, the results once again confirm the importance of these two non-sensory factors in the packaging design of mazamorra morada. The findings of this study highlight the importance of non-sensory factors that influence consumers’ purchase intention and acceptance of products, which can help stakeholders to design a package and a label that catches the attention of consumers.