Key Drivers of Consumption, Conceptual, Sensory, and Emotional Profiling of Cheeses Based on Origin and Consumer Familiarity: A Case Study of Local and Imported Cheeses in Greece

Abstract

The origin of a product, consumer familiarity, and purchasing identity are factors that affect the perception of cheese consumption. The present study aims at identifying consumers’ conceptualizations and attitudes towards local Greek cheeses of the North-Aegean Sea islands, such as Ladotyri, Graviera, Kasseri, Kaskavali, Melichloro, and Kalathaki, some of which have a Protected Designation of Origin (PDO) status, as opposed to cheeses of non-Greek origin, such as Cheddar, Regatto, and Gouda. Sensory and emotional attributes of local, local PDO, and imported cheeses, as well as drivers associated with consumers’ choice and acceptance above and beyond their sensory attributes, were studied using three methods: (a) flash profile to gain insight into the sensory positioning of products and description of samples; (b) qualitative analysis of focus groups to pinpoint consumer knowledge, preference, and consumption criteria; and (c) a new methodology for natural language processing and sentiment analysis of social media posts to determine consumer conceptualizations. Social media posts have proven to be a valuable source of linguistic and cultural data for cheeses. Local cheeses, including PDO products, were found to be linked to village life and family gatherings, home, tradition, and childhood memories, with saltiness and hardness being their main sensory attributes. Imported cheeses were linked to fast food, pizza, and snacking, with elasticity and gumminess as prominent sensory qualities. The main criteria for purchase were intended usage and versatility, taste and texture, price, and familiarity. The findings provide key sensory attributes, information about consumer purchasing criteria, and relevant vocabulary for the promotion of cheeses as agri-food and gastronomic identity key products.

1. Introduction

Gastrotourism is a type of cultural tourism with local food being an essential destination attribute, as production is a locally embedded activity which emphasizes cultural distinctiveness, authenticity, and sense of place, thereby facilitating competitive advantage for destinations and tourism businesses [1]. Local products in this context have additional value for tourists, often influencing pre-trip behavior, such as choice of destination [2]. Moreover, ethnocentrism and regiocentrism, that is, the tendency to consciously prefer products originating from the country or region one belongs to, as a means of supporting the local economy and part of consumer’s purchasing identity [3,4], attributes value to local products for locals as well. Product origin also determines consumer familiarity, which, combined with liking of the sensory attributes of foods and the emotional impact they have on the consumer, often alluding to memories, drives purchase [5,6].

For these reasons, a thorough understanding of how consumers perceive food products conceptually, sensorially, and emotionally is required for aligned product development and marketing. Although there is an increasing worldwide tendency to monitor consumers’ attitudes and behaviors, the effect of the country or region of origin of products on purchase and consumption is less extensive, especially in Greece [7,8].

The methods used in the present study comprise the emerging use of social media as a source of consumer-related data. Data sources such as online social media networking sites, content communities, online reviews, forums, and blogs provide a rich and expansive source of qualitative data that can be analyzed in a quantitative manner [9]. Especially when the focus of a consumer study is the phenomenon of food neophobia, data mining on social media platforms is preferable compared to common tests of central location to minimize selection bias [10] and gain access to more data. Age, gender, socioeconomic status, ethnicity, and country of residence seem to be important factors that are associated with social media usage [9], and therefore, they introduce bias that limits the generalization of the findings. However, because of the richness and amount of data they provide, social media platforms are currently used for language- and food-related research. Facebook, Instagram, and Twitter are mainly used, as they contain text and images or videos in posts by users and businesses, allowing for interaction between the account owner who posts and the account’s followers.

Food research in social media also utilizes sentiment analysis (SA) in search of consumers’ emotions towards products. Sentiment analysis or opinion mining is the computational study of people’s affective states (e.g., opinions, emotions, attitudes) toward entities, issues, events, topics, and their attributes [11]. In the case of food consumption, companies want to know consumers’ opinions about their products and brands. Sentiment classification is usually formulated as a supervised learning problem with three classes: positive, negative, and neutral.

The present study investigates how Greek consumers perceive cheese products and tries to answer questions such as: How are local cheeses perceived as opposed to non-Greek? What do the conceptual, sensorial, and emotional profiles of cheese products look like? What are the key drivers of cheese consumption? Do local products with a Protected Designation of Origin (PDO) label have extra value compared to non-PDO local cheeses? As mentioned in the literature, a PDO label creates expectations in consumers and elicits more intense positive emotions which the actual sensory experience often does not support [12]. Cheese is an integral part of Greek culinary culture; it is extensively produced throughout the country (252 thousand tons in 2023 [13]) and consumed as an appetizer to accompany wine or ouzo; as a side dish, in salads, pies, and pastries; as a pasta topping; as a recipe ingredient in casseroles, etc.; it can be practically consumed in any meal of the day (per capita consumption was 21.67 kg in 2022 [14]).

The insight gained from this study can be applied in gastrotourism marketing and for local cheese production alignment with consumer demands. In this study, a new methodology for data collection from social media and sentiment analysis using natural language processing tools and artificial intelligence applications is presented, alongside the convenient methods. More specifically, the proposed new methodology has been developed for the mining and handling of data on Instagram for specific language-related purposes, combining existing programming and natural language processing tools in an original way to decrease the need for manual data handling. Llama [15], a free open-source tool, is used for the automated application of categorization criteria as a substitute for humans to save time and ensure the repeatability of the results. An original formula is proposed for the measurement of hashtag engagement that is also applicable in other settings.

2. Materials and Methods

In total, a convenient sample of 51 Greek volunteers was considered in the study. Three methods were applied for data collection: flash profile, focus groups, and a social media platform. Flash profile methodology was used for conceptual and sensorial profiling of cheeses; focus groups were run for conceptual, sensorial, and emotional profiling of cheeses, as well as to identify key drivers of consumption and consumer perception of PDO products; and a popular social media platform was used to collect data on concepts and emotions relevant to cheese consumption. The cheese types selected for the study were local cheeses of the North Aegean, the prefecture of Greece where the studies with participants were performed, and imported cheeses corresponding to the others in texture and which are widely available in the area. Detailed information on the cheeses considered can be found in

Table 1. Cheese product types used as hashtags for collection of Instagram posts.

Local Greek Cheeses	Non-Greek Imported
PDO Ladotyri of Mytilene Mastelo® of Chios PDO Kalathaki of Lemnos PDO Feta of Lemnos Anthotyro of Lemnos PDO Kasseri of Mytilene Kefalotyri of Mytilene Melichloro of Lemnos Kathoura of Icaria Graviera of Mytilene Anthotyro of Samos Dry Mizithra of Mytilene Kaloupaki of Mytilene	Gouda Edam Regatto Mozzarella Emmental Parmesan Blue cheese, Roquefort, Gorgonzola Cheddar Cottage

. Not all cheeses were studied with all methods, as studies with consumers have a limit on the number of stimuli that can be presented.

2.1. Flash Profiling of Cheese Samples

Flash profile was conducted as described by Dairou and Sieffermann [16], according to whom a small number of untrained participants is sufficient to describe and rank samples. The process comprised three sessions in which the same 20 panelists (aged 20–33, median age: 22, 20% (n = 4) male) participated, with a briefing before each session. The small proportion of males is not a limitation, as according to the literature, females are more expressive and more elaborate with words, making them better subjects for consumer studies that require vocabulary elicitation [12,17]. Samples were presented blinded with three-digit codes in randomized order. Seven samples of cheeses were presented to the participants: Kaskavali of Lemnos, Melichloro of Lemnos, PDO Kasseri of Mytilene, Graviera of Mytilene, Regatto, Gouda, and Cheddar, with the cheese names written on labels. The cheese types were selected on the grounds of origin and texture. In Session 1, samples were presented simultaneously, and panelists were asked to list the sensory characteristics that best described the samples, avoiding hedonic terms (e.g., like, dislike, and pleasant). Each judge (panel member) created their own list of attributes. During Session 2, all lists created by the panel in the previous session were merged into one and presented to all judges, with the terms they had provided highlighted. The judges were then asked to individually proceed to rank the samples on a free-rating line for each attribute separately using the collective list. Session 3 was a repeat of the ranking task for repeatability reasons, and only repeated rankings were kept for further analysis. Each session lasted 20–30 min. Breaks were allowed and ties were permitted during ranking. Judges could evaluate and/or retaste the samples in any order, as many times as they needed. Generalized Procrustes Analysis (GPA) was applied for the consensus configuration between the judges’ sensory maps. The GPA plot demonstrates how similar or different the samples were based on the sensory profiles created by each participant. Data were handled in Microsoft Excel and analyzed with ANOVA, Spearman’s correlation test, and GPA, using XLSTAT as software (Addinsoft 2022.4.1.1358).

2.2. Focus Group Discussions on Cheese Consumption

Following standard focus group methodology, requiring 8–12 participants and a repetition of the procedure with at least three groups [18], 31 participants were randomly assigned to three groups (10, 10, 11) (aged 21–24, median: 22, 32% (n = 10) male). The small proportion of male participants is not a limitation, as explained in 2.1. The participants were asked to state their origin (place of birth and place where they had lived the most) using the 13 official prefectures of Greece. The discussion guide was prepared and executed by a member of the research team who was experienced in focus group moderating. The discussion consisted of questions regarding the frequency of cheese consumption, criteria for purchasing cheese, knowledge about the PDO label (body that assigns the label, process of assignment, what the label means for the product and the consumer), sellers that make them feel secure, instances of local and imported cheese consumption, and what descriptive words and emotions they correlate with local and imported cheeses (Supplementary Table S3a). Participants were also presented with 9 pictures of cheeses, namely, PDO Ladotyri of Mytilene, Melichloro of Lemnos, PDO Kalathaki of Lemnos, Graviera of Mytilene, Kaskavali of Lemnos, PDO Kasseri of Mytilene, Regatto, Gouda, and Cheddar, to be grouped after reaching a consensus. The cheese types were selected for the study on the grounds of origin, PDO status, and texture. No criteria for the number of groupings were imposed.

2.3. Data Collection from Social Media Cheese-Related Posts Using Natural Language Processing Tools and Artificial Intelligence Applications

Instagram was the social media platform preferred for this study because: (a) it facilitates multiform posts compared to Twitter, on which posts contain text only, and to Facebook, on which posts can be text only; (b) it has the highest penetration rate in Northern Europe as of January 2023 [19]; (c) it is the second most widely used platform by marketers worldwide (used by 80%) following Facebook (used by 89%) [20], and Instagram users interact with companies more often than on other platforms in Greece [21]; and (d) cooking/baking come third among the ten most common hobbies and second (together with health/fitness) of the most common interests of Greek Instagram users [21].

For data collection, Apify [22], a platform that provides a graphical user interface that allows developers to build, deploy, and monitor web scraping and browser automation tools, was used. Llama 3 [11], a large language model (LLM), was used for the automated application of the food-relatedness criterion and for sentiment analysis. The Llama 3 was chosen because: (a) it has the ability to process data in multiple languages, (b) it has the ability to process emoji/emoticons (tested on different types of posts in a preliminary study), (c) it is an openly available non-proprietary LLM (no cost), and (d) it is one of the top-performing non-proprietary LLMs in the LMSYS Chatbot Arena Leaderboard [23].

For this study, post engagement was measured. As there is no standardized method for post engagement measurement yet, and, to the extent of our knowledge, none for dealing with individual hashtags (a word or phrase preceded by the symbol # that classifies or categorizes the accompanying text (such as a tweet) [24]), a new method was developed. Following the model of Eriksson et al. [25], the engagement was calculated as the weighted average of the number of likes and of the number of comments, using 1 and 5 as the weights for the number of likes and comments, respectively. The sum of these two values is the engagement of the post, as shown in the following formula:

$$engagement=1\times \mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{l}\mathrm{i}\mathrm{k}\mathrm{e}\mathrm{s}+5\times \mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{s}$$

Since, for the present study, the focus was on hashtags, not posts, and hashtags appeared in posts of various engagement values, the normalized hashtag engagement was computed according to the following formula:

$$normalizedhashtagengagement={\displaystyle \frac{1\times {\left(\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{l}\mathrm{i}\mathrm{k}\mathrm{e}\mathrm{s}\right)}_{average}+5\times {\left(\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{s}\right)}_{average}}{{\mathrm{e}\mathrm{n}\mathrm{g}\mathrm{a}\mathrm{g}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}}_{median}}}$$

The median average engagement was then calculated. Finally, each engagement value was divided by the median (value/median), and the normalized value for each post expressed how many times the hashtag was more engaging than the average within the collected data. Posts with no likes nor comments were allocated a value of 0.0. The final value of each post was attributed to every hashtag of the post.

Posts on Instagram, a popular social media platform, containing the names of cheeses under study (Table 1) as hashtags in Greek and English were collected. The search was confined to hashtags in Greek for non-Greek cheeses to ensure focus on posts by Greeks, as the case study focused on Greek consumers (Supplementary Table S1). Meta’s Llama 3-8B model was used for sentiment analysis of the posts. It was prompted to check each post (caption, hashtags, emojis, comments) and assign a label, i.e., positive, neutral, or negative, in an automated process using the Python programming language [26] and libraries, e.g., Pandas [27]. Then, again using custom Python scripts, the frequency of appearance of the hashtags (i.e., how many times each hashtag appears in the posts collected) and the normalized average hashtag engagement were calculated.

3. Results

3.1. Flash Profiling of Cheese Products

During the first session of profiling the cheese samples, 97 unique descriptive terms were collected from participants in a pooled list, and synonymous terms were grouped together (Supplementary Table S2). During the following two sessions, samples were ranked regarding the vocabulary selected. The samples were analyzed in correlation to the rankings (18 participants’ responses qualified as consistent) for each word. The results of the GPA (p < 0.05) suggest that Kaskavali of Lemnos and Melichloro of Lemnos correlated with words such as intense, wine, salty, pictures, and granny’s cheese. Regatto slightly correlated with the same words. Graviera of Mytilene and PDO Kasseri of Mytilene correlated with words such as meze (appetizers), summer, and bread. Gouda correlated with words such as pizza, soft, gummy, night with friends, and sandwich. Cheddar correlated with words such as fast food and yellow. Kaskavali of Lemnos and Melichloro of Lemnos were very closely placed on the biplot. The same was true for Graviera of Mytilene and PDO Kasseri of Mytilene (Figure 1,

Table A1. F-values (p < 0.05 for ANOVA) and SCC values (p < 0.05) for Spearman’s correlation test on sensory attributes from each judge in the Flash profile.

Judge 1	F	SCC
sandwich	50.63 ***
gummy	2.608
white chocolate	1.255
earthy	3.377
fruity	2.500
granny’s cheese	124.83 ***	0.98 **
buttery	6.23 *	0.685
Judge 2
hard	18.75 ***	0.88 **
fatty texture	4.77 **	0.607
village	2.722
nights with friends	14.87 ***	0.85 **
Judge 3
fatty texture	6.70 **	0.704
soft	0.119
intense	15.94 ***	0.86 **
gummy	35.83 ***
soft	0.056
salty	6.38 **	0.691
Judge 4
pizza	18.40 **	0.88 **
plastic	60.66 ***	0.96 **
childhood memories	4.56 *	0.593
traditional	248.50 ***	0.99 **
Judge 5
sandwich	31.50 ***	0.929
village	35.833	0.937
fast food	5.093	0.627
Judge 6
intense	15.167	0.85 *
sweet	257.83 ***	0.99 **
fatty	5.15 *	0.631
traditional	24.26 ***	0.91 **
creamy	257.83 ***	0.99 **
salty	8.29 **	0.664
Judge 7
wine	14.07 **	0.85 *
intense	11.05 **	0.81 *
home	7.18 **	0.721
Judge 8
soft	8.42 **	0.757
smooth	5.83 *	0.667
sandwich	8.16 **	0.750
Judge 9
sandy texture	257.83 ***	0.99 **
granny’s cheese	18.39***	0.88 *
pictures	50.63 ***	0.95 **
nights with friends	6.09 *	0.679
Judge 10
salty	15.16 ***	0.85 *
family meal	15.16 ***	0.85 *
hard	31.50 ***	0.93 **
fast food	18.75 ***	0.88 *
intense	9.72 **	0.78 *
Judge 11
salty	31.50 ***	0.92 **
plastic	27.09 ***	0.92 *
buttery	20.22 ***	0.89 *
fatty texture	2.818
classic	253.16 ***	0.99 **
countryside	6.40 *	0.692
appetizers	9.52 **	0.782
family/home meal	49.70 ***	0.95 **
Judge 12
meal	64.16 ***	0.96 **
breadsticks	6.99 **	0.714
plastic	8.16 **	0.750
wine	11.90 **	0.82 *
fatty	9.72 **	0.78 *
Judge 13
salty	1.803
familiar	7.76 **	0.739
bread	11.16 **	0.81 *
tasty	3.417
sweet	9.19 **	0.775
soft	6.85 *	0.709
hard	5.15 *	0.631
gummy	14.87 ***	0.85 *
mom	64.16 ***	0.96 **
nights with friends	4.53 *	0.591
creamy	30.33 ***	0.92 **
childhood memories	8.16 **	0.750
traditional	11.90 **	0.82 *
rich texture	2.302
milk	2.926
fatty	2.818
intense	7.11 **	0.718
sour	6.10 *	0.679
classic	16.10 ***	0.86 *
Judge 14
intense	2.381
salty	4.58 *	0.595
soft	8.42 **	0.757
Judge 15
traditional	5.36 *	0.643
rich texture	15.16 ***	0.85 *
milk	6.10 *	0.679
cool	64.16 ***	0.96 **
countryside	2.112
holiday	3.542
spaghetti	8.16 **	0.750
childhood memories	3.079
tasty	1.398
Judge 16
spaghetti	10.72 **	0.80 *
sour milk	0.567
village		1.00 ***
childhood memories	248.50 ***	0.99 **
sandwich	7.15 **	0.720
breakfast	5.46*	0.648
Judge 17
salty	31.50 ***	0.92 **
spaghetti with cheese		1.00 ***
pizza	20.61 ***	0.89 *
sandwich	20.61 ***	0.89 *
hard	15.16 ***	0.85 *
Judge 18
happy memories	9.19 **	0.775
war	1.564
respect	4.34 *	0.577
yellow	6.38 *	0.691
home	18.75 ***	0.88 *
smooth	7.76 **	0.739
Judge 19
yellow		1.00 ***
soft	15.16 ***	0.85 *
salty	2.917
Judge 20
salty	31.50 ***	0.93 **
appetizers	50.63 ***	0.95 **
grandpa	7.18 **	0.721
spaghetti and mince	4.27 *	0.571
pizza	18.75 ***	0.88 *
smooth	0.353
summer	15.16 ***	0.85 *
village	22.37 ***	0.90 *
hard	11.90 **	0.82 *

For F values: * p < 0.05, ** p < 0.01, *** p < 0.001. For SCC values: * p < 0.05, ** p < 0.01, *** p < 0.001.

).

3.2. Focus Groups

The participants proved to be frequent consumers of cheese products, most consuming cheese every day or at least 4 to 5 times a week. They consumed local cheeses, especially Feta cheese, daily at home or when eating out at a tavern (traditional Greek restaurant) in salads, with bread, on the side, or with pasta, and consumed Kasseri in sandwiches. They consumed non-Greek cheeses daily, especially Gouda in sandwiches, Gouda or Mozzarella on homemade pizzas, Parmesan with pasta, and premium cheeses (smoked and spicy) with wine on special occasions, usually during social outings.

They did not usually look at the nutritional statement tables on the packaging as they usually purchased the same types or brands, and they were familiar with the amount of salt and fat that the products contain. If they did look at the table, they focused on protein, salt, fat, and energy in this order. Manner of consumption (intended usage and versatility) ranked first as the main key driver of cheese type selection, followed by the sensory attributes of taste and texture, as well as price. The saltiness (quantity of salt), origin of the cheese, familiarity with producer, environmental print of the producer, and packaging were mentioned among the top three drivers for purchasing. The company being environmentally friendly was not considered a virtue, as it was a minimum requirement for every company as mentioned by the participants. Packaging was important only regarding practicality, i.e., it must be convenient and able to protect the content after multiple openings. When asked about their preferred cheeses for emotional reasons, local cheeses of various parts of Greece were mentioned, such as Feta, Ladotyri of Mytilene, Mizithra, Kalathaki of Lemnos, Melichloro of Lemnos, and Graviera. The only cheeses of foreign origin mentioned were Mozzarella and Haloumi of Cyprus. When asked about their preferred cheeses for sensorial reasons, Feta came first again, followed by Cheddar, Graviera, Ladotyri of Mytilene, and Mozzarella.

When asked about the PDO status, process of status approval, and bodies involved, confusion and uncertainty were notable. They allocated a value of 2.5 out of 5 (allocated values by participants ranged from 1 to 4) to the importance of a cheese product being of PDO status. They stated that it is important for local cheeses to be based on traditional recipes and processes whether they have a PDO status or not, and that taste is more important than the labels. They also mentioned that they expect PDO cheeses to be more expensive because of PDO certification costs and for marketing reasons (i.e., that PDO is of higher quality). Regarding quality, it was not clear to the participants whether a PDO label actually indicated being of higher quality. They were skeptical and felt that it may be so at the beginning, but they stated that unvaryingly high quality is a requirement for PDO products. Regarding the sensory attributes of PDO products, they were considered to be consistent, but not necessarily tastier than non-PDO products.

When prompted to provide relevant vocabulary for their favorite non-Greek cheeses, participants used words such as: gummy, creamy, soft, salty, tasty, fragrant, wine, sensual, wintery, everyday, Italian, expensive, and pasta. For their favorite local Greek cheeses, they provided vocabulary such as: memories, tradition, classic, family, granny, habits, salty, hard, summer, goes with everything, and tasty.

The participants generally felt secure when purchasing cheese products. Purchasing packaged products from the supermarket was their first choice, followed by purchasing from a producer they know personally. Packaged supermarket products made them feel secure because they contain preservatives and are regularly inspected. Products directly purchased from a small-scale producer made them feel secure as well, as such products are expected to contain natural preservatives or no preservatives (Supplementary Table S3a).

The nine pictures of cheeses given to the participants of each group were put into three, three, and four groupings by the three focus groups (Supplementary Table S3b).

3.3. Social Media Cheese-Related Posts and Hashtags

The search retrieved 16,878 posts on local cheeses that contained 48,780 hashtags and 1158 posts on non-Greek cheeses that contained 6160 hashtags. The cheeses which were searched for were ranked in order of frequency of appearance according to the hashtags in the posts collected and in the order of the normalized average engagement of the respective posts (Figure 2). Graviera was both the most frequently appearing and the most engaging of the cheeses under study.

Sentiment analysis was performed using the Llama 3-8B model on the posts collected. Local cheese-related posts were identified as positive (90%), neutral (8.5%), and negative (1.3%), while non-Greek cheese-related posts were identified as positive (92%) and neutral (8%) only.

4. Discussion

Each method used in this study provided insight into consumers’ conceptualizations; the sensory and emotional profiles of local, PDO and non-PDO, and non-Greek cheeses; as well as the key drivers of their consumption. The findings were aggregated in a concept map (Figure 3). The main key driver of cheese consumption is intended usage, which is determined by texture (gumminess and hardness) and saltiness. In a study with Norwegian and French participants, appropriateness was the second most important purchasing criterion after price [28]. Price emerged as the leading key driver of purchase in a study with Spanish [29] and Portuguese consumers [30]. In the present study, texture and saltiness emerged as the most significant sensory attributes of cheese products, with local cheeses being perceived as hard and salty, while non-Greek cheeses were perceived as gummy and soft or creamy. Familiarity and price followed as drivers of purchasing, in line with the relevant literature [31,32], which is not surprising, as cheese is consumed almost daily in Greece and consumers usually buy the same products without having to overthink. Price, however, is a criterion that could alter habitual purchases. Saltiness and fat content are health-related attributes that emerge as important to cheese consumers, correlating with levels of education and knowledge about health [29].

Local cheeses seem to elicit a plethora of vocabulary in consumers’ minds, more than non-Greek cheeses, a result of extended past experience and a high level of familiarity [31]. They are conceptualized with childhood memories, family traditions and familiar faces, the countryside and village life, the Greek islands, summer, and other staple Greek foods. Non-Greek cheeses are conceptualized as more modern, alluding to foods closer to the Italian culture. Theses conceptualizations create a dichotomy between local and non-Greek cheeses in their emotional perception, with non-Greek cheeses eliciting emotions of sensuality and being financially privileged, while local cheeses elicit memories and a feeling of appreciation for tradition.

Most cheese types mentioned in the focus groups which were preferred for both sensorial and emotional reasons were identical, which suggests that both parameters are equally important in cheese consumption. This could be attributed to the tendency of individuals to consume cheeses with which they are familiar; which they have been consuming for years; and which elicit positive emotions, memories, and satisfaction [31]. Feta appeared to be the most popular cheese for both sensorial and emotional reasons according to the focus groups participants, while Graviera was found to be the most popular for Greek Instagram users. This finding supports the mention of Feta on the official website of the Greek National Tourism Organization as one of the most recognized Greek products [33].

The cheese groupings by flash profile analysis and by focus group participants point to the fact that local products are generally perceived as different than imported ones. The flash profile participants, who tasted samples, identified common attributes among the local cheeses, while focus group participants, who did not taste the cheeses and only had a picture, a name, and their experience, placed the local cheeses in different groupings. Intended usage related to texture (hardness and gumminess) and saltiness seemed to be the main criteria. The only local cheese considered similar to the non-Greek ones according to both the focus groups and flash profile analysis was PDO Kasseri of Mytilene, probably due to similarities in texture. PDO labelling did not seem to affect the groupings, which aligns with the focus group participants’ opinions shared during discussion that the label may indicate a higher quality and price, but the lack of it does not indicate poor quality. In the literature, consumers with higher educational backgrounds were willing to pay more for PDO-labeled and otherwise certified products [29,34].

Sentiment analysis of posts related to local and non-Greek cheeses did not provide a distinction between the two categories. Both categories are talked about positively to a significant degree on Instagram. This is in line with the literature, where it is stated that food is, in general, a positive experience for healthy humans [35], and that social media posts are more likely to be positive than negative, as positive posts elicit higher engagement [36]. Cheese rankings in terms of frequency of appearance and engagement showed that the most frequently posted about and engaging cheeses were all local ones, with Graviera being the most popular (most frequently posted about and engaging) cheese of all for Greek Instagram users.

A broader age range of participants could be pursued to test whether age is a parameter in cheese consumption, as there is conflicting evidence in the literature [37]. In addition, tourists, both Greeks visiting the area under study and non-Greeks, could be invited to participate in further studies for a better insight of what non-locals expect to experience. The findings could enlighten local producers as to what consumers expect from local cheeses, what makes them discrete, and how they could better promote products and modes of consumption.

5. Conclusions

Intended usage, which is determined by the sensory attributes of cheese, mainly texture and saltiness, seems to be the main driver of cheese consumption, followed by familiarity and price. The sensorial and emotional preferences for cheese seem to coincide. PDO status is not perceived as invariably linked to higher quality. Local cheeses, generally characterized as hard and salty, closely correlate with childhood memories and family meals, while non-Greek cheeses, generally characterized as gummy and soft, correlate with snacking and outings with friends. Relevant research could be extended to other local products, such as olive oil and wine, or products local to other areas of Greece and the Mediterranean, to identify similarities and differences nationally and cross-culturally. PDO products seem to require further study in terms of conceptualizations, emotions, and expectations to achieve better marketing and convince consumers, especially younger generations, that the label actually certifies quality and is worth the higher price.

The findings of the present study can also be used to create a gastronomic identity for the North Aegean islands after other studies are conducted on other local products. This gastronomic identity can, in turn, help to form the gastrotourism identity of the area [38]. Relevant concepts, sensory attributes, and elicited emotions can find their place on menus in local restaurants and hotels, as well as on the packaging of local products. Having a standardized and uniformly used concept map of local products, including but not limited to their sensory attributes, could help tourists choose the type they would expect to like according to their past experiences or identify one that will provide a new experience. Local cheeses can be used in modern recipes, substituting non-Greek cheeses with similar sensory attributes, an emerging trend in haute cuisine [39], since texture and saltiness seem to be the main focus of the consumer. If the agri-food economy, which seems to be declining in Greece, is linked to tourism, which is booming, it can have a prosperous future [1,40].