Understanding Australian Wine Consumers’ Preferences for Different Sparkling Wine Styles

: This study investigated the perceptions and preferences of Australian wine consumers towards different styles of sparkling wine, including French Champagne and Australian sparkling white, red and rosé wine, Moscato and Prosecco. An online survey of 1027 regular sparkling wine consumers captured demographic information, sparkling wine perceptions and preferences, and typical spending and consumption patterns. Consumers were segmented into three distinct clusters (‘No Frills’, ‘Aspirants’ and ‘Enthusiasts’) using the Fine Wine Instrument model. The majority of No Frills consumers were female and typically consumed sparkling wine once per month. Almost 55% of Aspirants were male with a household income of more than AU$75,000. Enthusiast consumers were also predominantly male and well educated, and 64% were under the age of 35 years. Sparkling white wine and Champagne were generally the preferred styles for each consumer group, followed by Moscato and sparkling rosé wine. Interestingly, Moscato scored favorably with both No Frills and Enthusiast segments. Almost 25% of respondents indicated that they were not familiar with Prosecco, while sparkling red wine was perceived similarly by male and female consumers. The findings from this study can be used by sparkling wine producers to better target their products and marketing to the specific needs and expectations of consumers within different segments of the Australian domestic market.


Introduction
When summarizing the key findings from wine consumer behavior research published over the past decade, Lockshin and Corsi highlighted the importance of researching premium and luxury wine behavior, successful marketing practices, and consumer behavior in emerging markets [1]. This included the value of wine tourism and marketing for value, as well as the relationship between grape and wine quality, and consumer behavior. Using market segmentation and a holistic approach to consumer behavior, a deeper understanding of consumer characteristics, habits, needs and expectations can be gained [2]. It is clear that a research gap exists with respect to classification of sparkling wine consumers and that categorization of consumers of this fine wine style would assist marketing strategies to target specific segments of the Australian domestic market. the country of origin [29]. However, gender was not found to have any impact on the magnitude of country of origin effects [29]. A subsequent study by Müller [30] distinguished six sparkling wine consumer groups, namely, the undemanding, the brand conscious, the ambitious, the region of origin conscious, the vine variety conscious and the experts. The study indicated that experts, the vine variety conscious and to a lesser extent the ambitious perceived the country of origin of the sparkling base wine to be important to their purchase decision making process and their willingness to pay [30]. Since reputation governs the preference order of all consumers, each purchaser will choose the product with the highest reputation he or she can afford [31]. It has been observed that individuals with higher internal values and more complex social identities were less susceptible to normative influence and placed less emphasis on social brand benefits [32]. Moreover, the most expensive and heavily advertised products are not automatically those preferred by regular wine consumers [33].
Rokka discussed how the image of Champagne has transformed from a practically insignificant no-brand wine label in the fifteenth century to an expression of modernity and icon for the global leisure class and celebration [34]. Australian contemporary counterparts made by traditional, Charmat, transfer and carbonation production methods have been shown to have varying chemical and sensory characteristics [35]. Segments of consumers of these sparkling wine styles, including Moscato, have disparate preferences for varietal and complex wines. Moreover, individual liking scores have enabled the identification of two consumer clusters with opposing preferences for distinct styles of Moscato [36]. Grape variety has a strong effect on the manufacture of sparkling wines, and those produced from innovative varieties (including Moscato Embrapa and Villenave) contain a high concentration of esters [37]. Additional research has also confirmed that the production method of Moscato Giallo wines (traditional, Charmat or Asti) influences the volatile composition of the sparkling products [38].
Previous research from Germany showed that Prosecco was mostly bought by people who preferred white wines or who did not have any preference for red wine [39]. An earlier Italian study examining the behavior of Prosecco consumers found that Controlled and Guaranteed Denomination of Origin (CGDO) consumers typically expressed a preference for CGDO products, and that they might be more loyal than Controlled Denomination of Origin (CDO) purchasers [40]. Italian consumers buying wine from supermarkets were surveyed, and substantial differences were observed amongst preferences for brand, certification of origin, and production practices (e.g., sparkling vs. semi-sparkling) [41]. A model to derive a reasonable pattern of differences in willingness to pay for Prosecco between CDO and Typical Geographic Indication types has also been developed [41]. Lastly, sparkling red is considered to be an iconic Australian wine style developed by French winemaker August D'Argent in 1881 for the Victorian Champagne Company [42]. However, Cohen and colleagues indicated that sparkling white wines have a higher likelihood of success and frequent purchase as opposed to sparkling red wines, which might indicate 'change of pace' tendencies [43].
Given the trend towards increased consumption of sparkling wine and Champagne, particularly at special occasions, further research is needed to better inform winemaking and marketing decisions to ensure industry meets the needs and expectations of different segments of the consumer market. This study therefore sought to understand Australian consumer awareness of and preferences for different sparkling wine styles, and the influence of occasion and price on consumption behavior, using the Fine Wine Instrument [44] to segment consumers according to their wine knowledge and purchasing behavior.

Consumer Survey
Themes identified from an extensive literature review were used to develop an online survey, which was administered nationally using SurveyMonkey™ (www.surveymonkey.com; San Mateo, CA, USA). Australian consumers (n = 1027) were recruited using a market research company (TKW Research Group, Seaford, Australia; www.surveytalk.com.au), with participants from a broad cross-section of states; prior to data collection, the survey was trialed by 10 South Australian consumers. Screening was performed using inclusion criteria that required participants to be at least 18 years of age and to have consumed sparkling wine at least 12 times per year on average. The survey took approximately 10-15 min to complete and data were collected over 2 weeks period. Participants were financially compensated for their time.
The questionnaire comprised three sections. The first section contained demographic questions relating to sex, age, education, and household income, as well as questions related to alcohol and wine consumption behavior. The second section measured fine wine behavior using the Fine Wine Instrument (FWI); a statistical model devised to segment consumers on the basis of wine connoisseur, knowledge and provenance variables [44]. Respondents were asked to indicate their level of agreement with a series of 18 statements using a 9 point category scale, where 1 = strongly disagree, 5 = neither agree nor disagree, and 9 = strongly agree. These statements were established in previous work by Johnson and Bastian [44] which investigated fine wine consumer involvement and identified three distinct types of consumers: 'Enthusiast' consumers, who exhibit connoisseur-like behavior, and are knowledgeable about wine and actively enjoy increasing that knowledge; 'Aspirant' consumers, who are less knowledgeable about wine, and less confident and adventurous in their wine-purchasing abilities; and 'No Frills' consumers, who display little connoisseur-type behavior, and who have little wine knowledge or involvement [44]. Section three of the survey examined participants' attitudes towards and preferences for different sparkling wine styles, specifically, Champagne, Australian sparkling white, red and rosé wines, Moscato and Prosecco. Consumers were made aware that sparkling wine should only be called Champagne if it originates from the region of Champagne in France. For the purposes of this study, all other sparkling wine styles (white, red and rosé, Moscato and Prosecco) were assumed to be Australian. Survey questions asked participants to list words that they associated with each of the sparkling wine styles, as well as any known brands. Respondents indicated their preferences for different styles using 9 point Likert scales (where 1 = extremely dislike to 9 = extremely like). Participants were also asked whether they would be likely to consume different sparkling wine styles at a number of pre-determined occasions (e.g., birthdays, Christmas, New Year and others identified in a previous study [45]), again using a 9 point category scale (where 1 = never, 5 = sometimes and 9 = always). Finally, participants were asked how much they would typically spend on a bottle of each style of sparkling wine at a retail outlet; with response options being: never purchase; <AU$15; AU$15-$29; AU$30-$49; AU$50-$79; and >AU$80.

Data Analysis
Consumer data were analyzed using a combination of descriptive techniques (frequencies, percentages, medians, means and quartiles) as well as agglomerative hierarchical clustering and nonparametric testing. Mood's median test [46] was used to test the equality of medians from two or more populations because the data are ordinal and the consumer segment responses did not follow a normal distribution. An examination of the interquartile ranges (IQR = 3rd quartile − 1st quartile) was also undertaken between the FWI segments for the different sparkling wine styles. The IQR is a measure of variability of FWI segment data (i.e., the spread of values), based on separation of a data set into quartiles. Fisher exact tests were used to test the association between qualitative variables given that some counts within contingency tables were less than 5. Statistical analyses were completed using XLSTAT 2016 (Addinsoft, New York, NY, USA). Qualitative analyses of word frequencies were performed using NVivo software Version 12 (QSR International Pty Ltd., Melbourne, Australia).

Influence of Consumer Segmentation on Preferences for Different Sparkling Wine Styles
Agglomerative hierarchical clustering using the questions developed by the FWI provided three distinct groups of wine consumers. As established by Johnson and Bastian [44], No Frills fine wine consumers demonstrate little connoisseur-type behavior, knowledge about wine or interest in the provenance of their wine purchases. The majority of consumers in this group were female (n = 287, 65.1%) and typically consumed sparkling wine only once per month (n = 256, 58.1%) ( Table 1). Aspirants share some similarities with the Enthusiast segment; however, their wine knowledge and wine involvement scores were all significantly lower. These respondents were not as confident in their wine-purchasing abilities and valued the opinions of others, including friends and family, staff at restaurants, wine retailers and wine writers [44]. Almost 55% of this segment were male (n = 266, 54.7%) with a household income of more than AU$75,000 (n = 289, 59.5%). Finally, Enthusiasts exhibit connoisseur-like behavior by keeping records of their wine purchases, having a special wine storage space and ritually checking their wines for faults prior to consumption [44]. These consumers were also mostly male (n = 62, 62%) and well educated (n = 66, 66% holding tertiary qualifications), and 64% (n = 64) were under 35 years of age. This was in agreement with Johnson and Bastian [44], who found a significant proportion of Enthusiasts were male and/or under the age of 35, and therefore highlighted the potential value in tailoring wine marketing strategies towards this demographic. Australian sparkling wine producers might similarly benefit from targeting a younger, male demographic, in order to better engage Enthusiast consumers.

Influence of Sparkling Wine Style on Consumer Perceptions and Preferences
Consumers were asked to list words and brands that they associated with each sparkling wine style, i.e., Champagne, sparkling white, red and rosé wines, Moscato and Prosecco. Forced open responses were collected, and participants could list as many or as few words/brands as desired. Similar to previous work undertaken by Verdonk and colleagues [47], word frequency analysis (including synonyms) was undertaken and is shown below, with results including word frequencies (i.e., the number of times each word appeared for each sparkling wine style) and weighted percentages for the top ten terms and brands ( Table 3).

Minimum | Mean | Median | Maximum
Consumers were asked to the list words and brands that they associated with each sparkling wine style (as many or as few words/brands as desired, but at least one response).
Of the 1027 regular Australian sparkling wine consumers surveyed, only 6.2% (n = 64) indicated that they were not familiar with Champagne. In contrast, only 10 respondents were unable to state their preferences for sparkling white wine (Table 5). Overall, 253 (24.63%) consumers were not familiar with Prosecco, 6.6% (n = 68) did not have an opinion about Moscato, and only 3.5% (n = 36) and 3.7% (n = 38) did not indicate a preference score for sparkling rosé and red wines, respectively. In vast contrast to the Enthusiast segment, the No Frills consumers demonstrated the least familiarity with Prosecco (n = 157), Champagne (n = 44), Moscato (n = 38), sparkling rosé (n = 26) and sparkling red wines (n = 25). Aspirant frequencies for all wine styles, except sparkling white wine (n = 5), sat between the No Frills and Enthusiast segments. In summary, the observed trend was that respondents were most familiar with sparkling white wine (n = 10) and least familiar with Prosecco (n = 253) ( Table 5).
The younger consumers who participated in this study (i.e., those under 35 years of age), preferred Moscato and sparkling rosé (medians = 7.0 for both, IQRs = 4.0, 2.0 respectively) more than consumers from other age groups. Respondents over 55 years of age (median = 8.0, IQR = 2.0) preferred sparkling white wine the most. Pairwise comparisons between all age categories for sparkling white wine and Moscato identified significant differences (p < 0.05). Nonetheless,   Data are the means, medians and quartiles of 9 point Likert scale scores (where 1 = extremely dislike, 5 = neither dislike nor like, 9 = extremely like, and 0 = never consumed). * p values at significance level of 0.05.    Data are the medians and quartiles of 9 point Likert scale (1 = extremely dislike, 5 = neither dislike nor like, and 9 = extremely like). * p values at significance level of 0.05. statistically significant differences were not observed between any of the age groups regarding sparkling red wine. Furthermore, consumers with postgraduate qualifications provided the highest scores for Prosecco (median = 6.0, IQR = 2.0). Whereas Moscato was most popular with respondents whose highest level of education was high school and trade qualifications (medians = 7.0 for both, IQRs = 3.0, 5.0). When comparing the median scores of all education segments collectively, Champagne (p < 0.0001), Prosecco (p = 0.001) and Moscato (p = 0.013) showed significantly different results. In addition, Champagne was the only sparkling wine style that provided significant differences between all income levels (p < 0.0001). A significant result (p < 0.001) was observed when comparing the lower Prosecco preferences of consumers who earn less than AU$50,000 to the higher scores of those who earn more than AU$150,000. Consumers with household incomes above AU$150,000 preferred Champagne the most (median = 8.0, IQR = 3.0), followed by sparkling white wine (median = 7.0, IQR = 2.0). The No Frills segment showed females preferred sparkling white wine (p = 0.002), Moscato (p = 0.002) and sparkling rosé (p = 0.033), significantly more than their male counterparts. Male and female Aspirant responses for all styles (except sparkling red wine) were significantly different; with females preferring sparkling white (p = 0.001), sparkling rosé (p < 0.0001) and Moscato (p = 0.002). The preference scores of male and female consumers in the Enthusiast segment were not significantly different for any of the sparkling wine styles (Table S1 (Supplementary Materials)). When comparing the preference scores of individual age groups (i.e., <35 years, 35-55 years, >55 years) within the No Frills segment, significant differences were observed between all the age groups for sparkling white (p = 0.004).
Significant differences were found when comparing the No Frills and Aspirant consumers' preferences for Champagne, according to those who had completed High School with those who had undergraduate (p = 0.025, p = 0.048 respectively) and postgraduate (p = 0.016, p = 0.001 respectively) qualifications. Aspirant respondents who had completed postgraduate study also provided significantly higher preference scores for Champagne than participants who had completed a trade qualification (p = 0.014). Preferences for Prosecco were higher from those who had finished postgraduate study compared to a trade qualification, for both the No Frills and Aspirant segments (p = 0.041, p = 0.019 respectively). Enthusiasts who had completed undergraduate education gave lower preference scores for Moscato when compared to High School and Trade School graduates (p = 0.001 for both) and Postgraduates (p = 0.011). A significant difference between Enthusiast preferences for Prosecco was also perceived amongst those who had been educated at undergraduate and postgraduate levels, where postgraduates preferred the style more (p = 0.041) (Table S3 (Supplementary Materials)).
Preferences of No Frills consumers of different income levels showed significant differences for sparkling white wine (p = 0.012) and Prosecco (p = 0.019). Those with incomes less than AU$50,000 were significantly different to those earning AU$50,000-$100,000 and greater than AU$150,000 (p = 0.019, p = 0.017 respectively for sparkling wine and p = 0.007 and p = 0.005 for Prosecco). In addition, significant differences were observed between the No Frills sparkling white wine preference scores of people who earned AU$100,001-$150,000 and AU$50,000-$100,000 (p = 0.031) and more than AU$150,000 (p = 0.021) (Table S4 (Supplementary Materials)).
When comparing preferences for Champagne across all segments of varying incomes, statistically significant results were observed (p < 0.001). In addition, within the Aspirant segment there were significantly different results across all income brackets (p < 0.0001). Mood's test showed that the median Champagne scores were different when comparing Aspirants with incomes less than AU$50,000 to AU$50,000-$100,000 (p = 0.004), AU$100,001-$150,000 (p < 0.001) and greater than AU$150,000 (p < 0.0001). When considering Champagne, a significant difference was also observed for Aspirants earning AU$50,000-$100,000 and those with a household income over AU$150,000 (p = 0.004). Aspirants with a household income of less than AU$50,000 had significantly different preference scores for sparkling red wine to those earning AU$50,000-$100,000 (p = 0.006). Finally, the Champagne preference scores of Enthusiasts who earn AU$50,000-$100,000 were significantly lower than those who earn more than AU$150,000 (p = 0.039). Enthusiasts with an average household income less than AU$50,000 provided significantly higher scores for sparkling rosé wine, than those who earn AU$100,001-$150,000 (p = 0.021) (Table S4 (Supplementary Materials)).

Influence of Occasion on the Consumption of Different Sparkling Wine Styles
The 'situational purchase context' is a principal driver behind sparkling wine purchasing [13] and Champagne has been described as 'the celebration wine' [10] which Australian consumers typically purchase with the intention of sacralizing events [51]. Anchor and Lacinova found that the second biggest motivation for drinking wine, especially sec or demi sec sparkling wine, was 'to celebrate something' [52]. It has been argued that a number of variables are affected by this situational context, including the country of origin effect, the price consumers are willing to pay and perceptions of prestige and luxury [13]. In the current study, when survey participants were asked whether they would consume different sparkling wine styles at a number of pre-determined occasions (identified during focus groups previously conducted by Verdonk and colleagues [45]), the results showed highly significant differences between the ranked medians of the FWI segments for all occasions (p < 0.05). Each of the occasions specified showed an increase in the likelihood of consumption as consumer involvement increased (No Frills median ≤ Aspirant median ≤ Enthusiast median). As argued by Spawton [53], the association of sparkling wine with celebration is a key reason why this style is chosen in preference to other alcoholic beverages. This was supported in focus groups held by Olsen, which revealed that participants perceived sparkling wine to be most appropriate for celebrations [54], and a 2016 study found that Croatian sparkling wine consumers generally associated consumption with specific celebrations [6].
In this study, the Enthusiast segment was most likely to consume every style of sparkling wine at each of the listed occasions (median ≥ 2, i.e., anniversary, at home with food, at home without food, birthday, breakfast, by yourself, Christmas, during the week, funeral, girl's/boy's night out, hot weather, Melbourne Cup, New Year, on the weekend, pub/club, restaurant/café, wedding, work drinks). In fact, the median numbers were above 4 for all wine styles, except Prosecco. These consumers were most likely to drink Champagne, sparkling white, red and rosé wines, and Moscato (median ≥ 5) at an anniversary, at home with food, birthday, Christmas, during the week, Melbourne Cup, New Year, on the weekend, pub/club, restaurant/café, wedding and work drinks ( Table 7).
The No Frills segment did not report consuming sparkling red, rosé, Moscato or Prosecco (median = 1) at an anniversary, at home with food, at home without food, birthday, breakfast, by yourself, during the week, funeral, girl's/boy's night out, hot weather, Melbourne Cup, New Year, on the weekend, pub/club, restaurant/café, wedding and work drinks. However, the Aspirant group was more likely to consume sparkling red and sparkling rosé (median ≥ 2) at an anniversary, at home with food, birthday, Christmas, New Year, on the weekend, restaurant/café and wedding. Of all the sparkling wine styles, Champagne and sparkling white wine were consumed the most at all occasions, with Enthusiasts consistently providing the highest scores. The highest median values for all segments consuming Champagne and sparkling white wine (i.e., No Frills median > 4.0, Aspirant median > 5.0, Enthusiast median > 6.0) were found on anniversaries, birthdays, Christmas, Melbourne Cup, New Year and weddings (p < 0.001 for all). Sparkling white wine was also regularly consumed on weekends (No Frills median = 4.0, Aspirant median = 5.0, Enthusiast median = 6.0) (p < 0.001) ( Table 7).
Pairwise comparisons between the consumer segments showed that the majority (96.5% of combinations tested) of relationships were significantly different. However, statistically significant

Influence of Price on Consumer Purchasing Behavior
Several studies have identified price as being an important consideration during wine purchasing decisions [53,[55][56][57][58][59][60][61], with high prices being associated with superior quality [53,59]. Six attributes were found to be statistically important in explaining deviations from average wine prices: quality, cellar potential, grape variety/style, region, vintage and producer size [62]. Lecocq and Visser found that price differences could be explained by characteristics which were directly revealed to the consumer upon inspection of the bottle and its label (ranking, vintage and appellation), rather than sensory variables [63].
All consumer segments were willing to pay more for Champagne than any other style of sparkling wine. This likely reflects the influence of country of origin and price on consumer perceptions of wine quality [64,65]. Evidence also suggests purchasers are willing to spend more per bottle when wine is purchased for special occasions [49]. Although not specific to sparkling wine, it has been suggested that associating a given wine with an occasion might assist consumers with their purchasing decisions [66].

Conclusions
Different styles of sparkling wine (both fruit driven and complex styles) appeal to different segments of the domestic sparkling wine market. In the current study, sparkling white wine and Champagne were the preferred wine styles, followed by Moscato and sparkling rosé wine. However, preference scores for sparkling white and rosé wines were significantly higher for women, than for men, and younger consumers (i.e., those < 35 years of age) preferred Moscato and sparkling rosé more than consumers from other age groups. Men and women liked sparkling red wine equally and Moscato appealed to both No Frills and Enthusiast consumers. Whereas Italian sparkling wines have enjoyed considerable (international) growth in recent years, most of the Australian consumers surveyed did not consume it regularly. In fact, almost 25% of consumers were unfamiliar with the style, suggesting Australian wine producers might benefit from further marketing this style. Perhaps not surprisingly, Enthusiasts consumed all sparkling wine styles, more often, and at different occasions, and were willing to spend more on Champagne, albeit, on average, the majority of respondents do not pay more than AU$50 per bottle for Australian sparkling wine. The outcomes of this study can be used by sparkling wine producers to better tailor their products and marketing strategies to the specific needs and expectations of consumers within different segments of the Australian market. This research aimed to address a knowledge gap regarding the categorization of sparkling wine consumers to assist marketers in targeting specific segments of the Australian domestic market. There are several limitations to this study, due to possible sample and self-selection biases of survey respondents. Despite a recruited convenience sample of approximately 1000 Australian regular sparkling wine consumers, it should be acknowledged that the participants may not be entirely representative of the broader Australian sparkling wine consuming population. In addition, the survey required participants to self-report data, which could also lead to accuracy issues. Opportunities for future research include consumer tastings to determine sparkling wine preferences, as well as an exploration of consumers' knowledge of sparkling wine production.
Finally, this study could be replicated in other countries, to determine how cultural influences affect consumer behavior.
Supplementary Materials: The following are available online at https://www.mdpi.com/2306-5710/6/1/14/s1. Table S1: Influence of FWI segmentation and gender on preferences for different sparkling wine styles.; Table  S2: Influence of FWI segmentation and age on preferences for different sparkling wine styles; Table S3: Influence of FWI segmentation and education on preferences for different sparkling wine styles; Table S4: Influence of FWI segmentation and household income (AUD) on preferences for different sparkling wine styles; Table S5: Influence of FWI segmentation on consumption occasions of different sparkling wine styles.
Author Contributions: All authors were involved in conceptualization; methodology and formal analysis, N.V.; writing-original draft preparation, N.V. and K.W.; writing-review and editing, R.R., J.C. and K.P.; supervision, K.W., R.R., J.C. and K.P.; funding, N.V. and K.W. All authors have read and agreed to the published version of the manuscript.