A recent study conducted by De Luca et al. (2019) served as principal inspiration for the present research. In this study it was assessed how consumers perceived aspects of dietary cookies when served in an unlabeled transparent-round-glass jar (as in a cookie jar), as compared to when served it its commercial/original packaging (in this case, an opaque bag which included labeling with branding and formulation; cf. Adam and Ali, 2014). These results, based on 31 European participants, suggested that such differences in packaging appearance may prompt different expectations on dietary cookies. For instance, this study reported that dietary cookies were thought as potentially lighter, but not necessarily healthier, when being offered in their original labeled packaging, as compared to when offered in the unlabeled transparent jar. Moreover, the participants rated the same cookies as potentially more caloric when presented in the unlabeled jar, as compared to the ratings under the influence of the original packaging. The obtained evidence also suggested that ‘lightness’ and ‘calories’ related wording were more effective message carriers for dietary cookies labeling, rather than ‘healthy’ type of wording, or wording related to flavor/hedonic sensations elicited by the cookies experience (i.e., crumbliness, greediness).

Based on De Luca et al.’s (2019) assessment, we decided to design this new study focused on carefully evaluating the effects of packaging transparency and type on expectations in an experiment resembling e-commerce. The latter is also aligned with the e-commerce growing trend for this food category (e.g., Market Analysis Report, 2019). As such, our new study would be principally about the expectations elicited by the digital visual appearance of packaging, where tasting the cookies and touching the packaging would not be part of our experimental scope (cf. Citrin et al., 2003; Overmars and Poels, 2015; Petit et al., 2019).

Here, we also saw the need to use customized brand-free labeling (something that was not considered in the aforesaid study). In fact, while designing our own experimental label for this new study, we would be able to optimize the amount and type of information present in such labeling (cf. Carrillo et al., 2012). The latter is relevant when assuming that the usual e-commerce platform does not always allow the consumer to properly read all the information that may be present in a dietary product’s label. Finally, we wanted to more accurately control the three dimensions of a packaging appearance that were explored in the latter study as well (jar vs. bag, transparency vs. opaqueness, unlabeled vs. labeled packaging). For this, we decided to compare such dimensions not only between two experimental conditions (as in De Luca et al., 2019), but across eight different versions of packaging.

The objective of this study, therefore, was to look for ways to customize the online experience of dietary cookies more accurately via its packaging. For this, we assessed how different combinations of three key aspects of a packaging appearance (with vs. without labeling, bag vs. jar, transparency vs. opaqueness) would affect the expectations of consumers while evaluating dietary cookies in a 2D digital context.

First, we hypothesized that an optimized label mainly communicating the dietary attributes on the packaging would be more strongly associated with healthier type of cookies, when compared to an unlabeled version of the same packaging. Second, we thought of comparing two packaging types (a cookie jar and a bag). Specifically, we thought that a jar usually has rounder shapes than most cookie bags (Robertson, 2011; Davidson, 2018). Based, for instance, on crossmodal correspondences research – which suggests that round shapes are more naturally associated with sweet/smooth sensations, when compared to more angular shapes (cf. Becker et al., 2011; see also Reinoso Carvalho et al., 2017b, for an example), one may expect that a jar may prompt higher expectations toward a sweeter, and/or smoother, type of cookie, when compared to a bag. Third, a transparent packaging would allow the cookies to be fully exposed, and where we hypothesized that this would have positive consequences on the expectations of consumers while evaluating the product (Petit et al., 2021; Simmonds and Spence, 2017; Simmonds et al., 2018). The latter may be even more relevant in a digital environment, where none of the senses, besides vision, tend to be significantly activated (at least not yet).

In brief, by contrasting the appearance of a packaging across such three dimensions, we tested the following main general hypothesis:

The test of the general main hypothesis was carried out across the following three specific hypothesis, and where each of these particular hypotheses focused on the three specific aspects of the packaging being manipulated:

In general, we assumed that the said modulation in expectations would be measurable in terms of the ratings related to healthiness (e.g., lighter, healthier, caloric content), differences in hedonic/qualitative ratings (e.g., tastiness, greediness – as in ‘these cookies look so good they make me hungry,’ price), and/or differences in sensory flavor ratings (e.g., sweetness, crumbliness).

In total, 496² participants took part in the experiment (n = 496; 62.50% males, Mean of age 25.71 years SD = 7.91). The survey was programmed on Qualtrics software³ and the participants were recruited on Prolific Academic⁴, in exchange for GBP 0.63. The focus on this recruitment was to model the general population, so no specific filters were set concerning the attitudes that the sampled population might have toward cookies or dietary cookies. Here, most participants reported eating cookies (95.00%), loving cookies (86.10%), and usually paying attention to food/drink product’s labeling while shopping (73.80%). Only 31.50% of the participants reported actually consuming light and/or dietary type of cookies. Concerning cookies eating habits, the obtained sample appears as representative of an European population. For instance, previous evidence has reported that 96% of such consumers tend to buy into this category, and that cookies with certain dietary characteristics are being consumed by more or less a third of this kind of population (Harris Interactive Survey, 2019).

Eight different versions of packaging were produced for this experiment. They were carefully designed with respect to any unintended influences, for example, when it comes to the quality prompted by the packaging type, as well as while keeping away from resembling existing cookies brands. Such eight different versions of packaging included two variations of packaging type (jar vs. bag), two variations of packaging transparency (transparent vs. opaque), and two variations of labeling (labeled vs. unlabeled). Based on Carrillo et al. (2012) and De Luca et al. (2019), the label included the product’s title (in this case, ‘Light cookies’), along with highlights on the product’s formula (‘low in calories’; ‘palm-oil, milk, and eggs free’). Figure 1 shows the different versions of the packaging stimuli.

This study followed a 2 (packaging type: jar vs. bag) × 2 (labeling: label vs. no label) × 2 (transparency: opaque vs. transparent) between-participants experimental design. Each participant was randomly assigned to, and asked to evaluate, only one of the eight versions of the dietary cookies packaging (see Table 1). In particular, they were asked to evaluate the packages in terms of healthiness, quality, greediness, price, lightness, calories, crumbliness, sweetness, tastiness, cookies/packaging liking. The completion of the survey lasted for approximately 5 min in total.

The packages were presented scaled to match in size (1041 × 1476 pixels), and large enough in order for the participants be able to read all the available information present in the label (the latter only when applicable, since four of the eight versions of packaging were unlabeled; see Figure 1). Since the transparent packaging allowed the participants to actually see the cookies, and the opaque did not, it was decided to include mini replicas of the cookies as part of the label.

The answers concerning the expectations elicited by the packaging were all based on 5-point rating scales (anchored with 1 = ‘Not at all’, 5 = ‘Very much’). After answering these questions, the participants were also asked to provide some demographic information (age, gender), and their general cookies-consumption habits. Appendix Table A, shows a summary of the variables assessed in the survey. The order of the presentations of the survey’s questions, and multiple-choice answers, were fully randomized.

A 2 × 2 × 2 multivariate analysis of variance (MANOVA) was conducted via SPSS 26, with type of packaging, labeling, and transparency as between-factors, and the rating-scales as the dependent variables (see Appendix Table A, for an overview on these dependent variables). Gender and age were included as controls (covariates).

Table 5 reports mean differences when comparing general presence vs. absence of label, and regardless of packaging type (jar vs. bag), and level of transparency (transparent vs. opaque). The variables being here considered are those which showed a significant difference in Table 4. This evidence points toward the presence of label in the packaging as triggering higher ratings in all of the assessed dimensions (except for high in calories, which had an opposite direction).

Table 6 shows the mean differences when comparing general transparency vs. opaqueness, and regardless of packaging type (jar vs. bag) and labeling (presence vs. absence). The variables being considered are those that prompted a significant difference in Table 4. Here, the evidence shows that a transparent packaging prompted higher ratings for quality, greediness, high in calories, crumbliness, sweetness, tastiness, cookies, and packaging liking ratings. On the other hand, an opaque packaging prompted higher ratings for healthiness, lightness, and low calories.

As shown in Tables 3, 4, there was evidence of a significant interaction between labeling and transparency on quality, greediness, high/low in calories, crumbliness, sweetness, tastiness, cookies, and packaging liking ratings (Figure 3). In particular, the ratings of quality, greediness, low in calories, and packaging liking, tend to decrease without labeling, and regardless of packaging transparency. Comparing to the latter, high in calories ratings have an opposite direction of effects. For crumbliness, tastiness, and cookies liking ratings, the ratings tend to decrease without labeling, but only for the opaque version of packaging. Concerning sweetness, evidence shows that the label tends to trigger similar effects on transparent and opaque packaging. However, the expected sweetness ratings tend to be enhanced for unlabeled transparent packaging and diminished in the unlabeled opaque one.

Tables 3, 4 show a significant interaction between type and transparency in packaging on quality, greediness, expensive, crumbliness, tastiness, cookies, and packaging liking ratings (see Figure 4). In particular, quality, greediness, crumbliness, tastiness, cookies, and packaging liking ratings tend to increase in the transparent, and decrease in the opaque, jar conditions. For the bag, however, ratings did not seem to be as significantly affected by changes in transparency, when compared to the jar. Concerning expensiveness ratings, people tend to expect the cookies as more expensive when presented in the transparent jar, as compared to the transparent bag. In addition, expensiveness ratings did not seem to be so significantly affected in the opaque version of both packaging.

Tables 3, 4 show evidence of a significant interaction between labeling^∗type^∗transparency, concentrated on healthiness, expensive, lightness, tastiness, and cookies liking ratings. Since no significant effects were reported in Table 3 for labeling^∗type, we decided to independently assess the moderating effect of type of packaging on labeling^∗transparency, and the moderating effect of labeling on type^∗transparency of packaging (Gignac, 2019).

The obtained results suggest that changes in the visual appearance of packaging, in terms of presence (vs. absence) of label, packaging type (jar vs. bag), and transparency (vs. opaqueness), modulate the expectations of dietary cookies in a digital environment (H1). For instance, concerning the main effects of each visual cue (labeling, type of packaging, transparency), the presence vs. absence of label in the packaging, as well as the ability, or not, of visually inspecting the cookies, had a significant effect on most of the ratings (e.g., H1a, H1c). Moreover, when it comes to the interaction effects, for example, there were moderating effects of type of packing on transparency, and on labeling^∗transparency (H1b).

Note that, when the participants were asked if they were consumers of dietary cookies, only 35.1% answered that they actually were (n = 174). Therefore, we decided to split the original dataset across those that reported consuming such types of cookies, and those who reported not doing so, and re-run two independent analyses. However, after running such analyses, no remarkable differences in the data were observed suggesting that such contrast in consumption habits triggered robust differences across the obtained results (cf. Lwin et al., 2014). Such null effects may also be related to the split of the sample and, hence, a consequent underpowered analysis.

As expected, the presence of labeling in the packaging primed the consumers toward a more precise understanding of the dietary characteristics of such cookies, when compared to unlabeled versions of the same packaging [De Luca et al., 2019; see section “Main Effect of Packaging Labeling (With vs. Without Label)”]. Hence, the short and precise information present in this experimental and customized label was able to effectively nudge the consumer toward the understanding of the principal dietary characteristics of the product at stake (i.e., Carrillo et al., 2012). What is more, even when the label did not specify any precisions regarding the sensory characteristics of the cookies, the presence of label also enhanced the ratings related to greediness, crumbliness, sweetness, tastiness, cookies, and packaging liking (cf. Pinto et al., 2017). Note that the cookies with labeling were also rated as the most expensive. Hence, we could further hypothesize that the presence of label gave support to a more meticulous framing of the cookies within a specific category (i.e., dietetic food), where prices are usually above the average when compared to the generic version of the same product.

There was no evidence of a main effect of type of packaging on expectations. However, while assessing the interactions, there were moderating effects of type of packing on transparency, and on labeling^∗transparency. For instance, as shown in section “Interaction Effect Between Type and Transparency,” people tend to expect cookies to be more expensive when presented in a transparent jar, as compared to a transparent bag. Moreover, bag ratings seem to be generally less affected by changes in transparency, when compared to the jar. Hence, and even though type of packaging did not prompt as robust main effects, when compared to labeling and transparency, it can still be of importance to consider the type of packaging in order to get more precise effects out of these other two visual cues, and especially when it comes to combining type of packaging with transparency.

The results show that a transparent vs. opaque packaging can generally trigger higher expectations concerning the overall quality (which is consistent with Sabri et al., 2020), but not necessarily when it comes to expected healthiness, [see section “Main Effect of Packaging Transparency (Transparency vs. Opaqueness)”]. In particular, we saw that the transparent packaging generally prompted higher expectations with concerns to aspects that could be understood as qualitative (think of a product’s quality, liking for the product and packaging, greediness), as well as sensory ones (crumbliness, sweetness). On the other hand, however, the transparency of the packaging had an opposite effect when it comes to the perceived healthiness, where the cookies were expected to be less light, and higher in calories, when presented in a transparent (vs. opaque) packaging. In fact, when considering the interaction effects reported in “Interaction Between Labeling and Transparency,” the unlabeled and transparent packaging was expected to provide sweetest cookies when compared to the unlabeled, but opaque, presentation. Hence, we could conclude that allowing the visual inspection of the cookies can encourage the consumer toward more hedonic – thus, potentially less rational – choices during decision-making tasks related to the online consumption of dietary food (Simmonds et al., 2018), and that this might be independent of the information that the brand can deliver to the consumer via labeling (cf. Adam and Ali, 2014). The latter may also be in line with the fact that transparent packaging can lead to better imagery, thus eating simulation could be the mechanism explaining such an effect (Petit et al., 2021).

A thoroughly customized packaging experience can have important consequences on the way a consumer interacts with a product. Based on the particular results of this study, it may be possible to more effectively nudge consumers toward healthier choices of cookies, while at the same time allowing brands of dietary cookies to better market such type of products. For instance, as shown in this study, if a brand of cookies is looking for to emphasize its dietary characteristics, they might as well prioritize communicating such attributes via effective labeling [as in section “Main Effect of Packaging Labeling (With vs. Without Label)”], and perhaps not rely on transparent packaging. However, another brand that may be more interested in evoking, say, greediness during decision-making tasks, could perhaps focus on the design of the cookies, and show them via transparent packaging [as in section “Main Effect of Packaging Transparency (Transparency vs. Opaqueness)”].

The obtained results are certainly relevant in a digital environment where vision (and to a lesser extent audition and touch), tends to be the principal sense (Petit et al., 2019). As a matter of fact, a profound customization of the visual aspects of packaging may be a good start when considering the future impact of new technologies on the stages of the consumer’s journey where, for instance, further advances in virtual/augmented reality will certainly transform the online shopping experience (Hoyer et al., 2020).

From a marketing perspective, this study also highlights how a food company – in an online environment – can work on its own identity by strengthen packaging communication. For instance, the obtained evidence offers packaging design solutions for safeguarding brand awareness, along with the possibility of enhancing quality and healthiness expectations during e-commerce decision-making tasks.

Concerning the study’s limitations, only visual aspects of packaging were tested. Therefore, future similar studies could compare, not only packaging aspects, but also different appearance of cookies. When it comes to the particular visual aspects of packaging that were tested, future studies could further manipulate packaging material, where sensations related to smoothness vs. roughness could be further enhanced (e.g., while retesting H1b). In this study we only tested presence vs. absence of customized labeling. However, the customization of labeling in dietary cookies could also be further addressed in similar future studies by, e.g., testing different combinations of messages along with particular color and/or font customization. Here we only tested one type of transparency in packaging as well. It may be of use to compare in the future glossy vs. matte transparent cookie jars, as it has been suggested that such differences can affect the perception of a packaging in terms of, e.g., haptics (Decré and Cloonan, 2019). Important to note as well that this study has the particular limitation of not considering that the choice of a packaging’s type is not only about aesthetics, but it might also be related to the performance of such packaging across the corresponding supply chain.

In brief, and even though this assessment has the abovementioned limitations, this study invites toward continuing exploring the further customization of packaging for dietary product categories, and especially considering potential new consumers of these type of products in digital environments.