1. Introduction
Market segmentation has been used as a strategy by companies since the 1950s to identify homogeneous groups of consumers within a heterogeneous market and thereby provide a competitive advantage, such as tailored products and marketing communications that meet the needs of the identified consumer segments [
1,
2,
3,
4]. The Australian wine market has a long history of academic study into market segmentation (see, for example [
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15,
16,
17]).
However, it has been a continuing challenge to identify the most suitable segmentation basis for this heterogeneous market [
18] as traditional approaches, such as those based on demographics, have been questioned due to a lack of richness [
19,
20,
21]. In recent years, “domain-specific” lifestyle segmentation (i.e., consumer segmentation based on psychographic data specific to a specific consumption situation or a set of consumption behaviours) has been identified as a promising segmentation method for food in general [
21] and specifically for wine (for example [
14,
15,
22,
23]). For more detailed analyses of the history of wine market segmentation and the various segmentation bases available, see Johnson and Bastian [
8] and Johnson, Danner and Bastian [
24,
25].
One such lifestyle segmentation study was that of Johnson and Bastian [
8], who argued that a multi-dimensional psychographic segmentation tool based on Attitudes, Interest and Opinions (AIO) would provide a more detailed and sophisticated description of any identified segments, than more traditional segmentation bases. They developed the Fine Wine Instrument (FWI) that consists of 18 scale items covering three dimensions, labelled connoisseur, knowledge and provenance. These three dimensions then formed the segmentation base for a cluster analysis that identified and characterised three fine wine-related segments in the Australian wine market.
Although there are numerous studies applying “domain-specific” lifestyle segmentation in different markets [
24,
25,
26,
27], all published studies solely relied on respondents’ self-reported answers collected through online surveys or face to face questionnaires. None of these studies investigated consumers within real life tasting situations nor how their responses differed by wine-related behaviour-based segments when actually tasting wines. The aim of the current research was to investigate consumer behaviour when tasting wines in different consumption contexts, consistent with the opinions of Jaeger et al. [
28] and Meiselman [
29], who advocated for research in more complex consumption situations as means to improve ecological validity over an unnatural sensory laboratory setting.
While product aroma and flavour characteristics are essential sensory elements underpinning consumers’ preferences and purchase decisions [
30,
31], consumer perceptions, liking and decision making are also influenced by the context of consumption [
32,
33,
34,
35]. Context considers social and environmental factors plus the product itself, including not only intrinsic attributes but also extrinsic characteristics such as the product related information provided at consumption [
34]. It has been shown that both experts and novice consumers rely on extrinsic cues when evaluating wine [
36] and that the consumption situation (different locations, wine with or without food) can influence consumers’ liking [
37,
38] and more general perception [
35] of wines.
The product information available to the consumer when tasting or consuming a wine, e.g., that presented on wine labels, can also be regarded as a contextual factor and has the potential to influence their expectations [
39] and hedonic responses [
40,
41]. D’Alessandro and Pecotich [
36] showed that when choosing, and judging wine quality and price, both wine experts and novices were influenced by extrinsic cues. However, little research has been published investigating context effects on different wine consumer segments as opposed to a heterogeneous consumer cohort, resulting in the first research question: Do the hedonic responses of fine wine segments (FWS) generated by the FWI differ when tasting wines of diverse varieties or quality levels in different consumption contexts?
In the past decade, research has shown that hedonic responses (i.e., liking) are only one component of the wine consumption experience and that consumption evoked emotions can provide additional insights into consumers’ perceptions and product related behaviour [
42,
43,
44,
45]. More recently Calvo-Porral et al. [
46] used consumers’ emotional associations with wine consumption as a segmentation basis and identified four distinct clusters: the unattached, the negatives, the circumspect and wine lovers. This raises the second research question: Do reported wine-evoked emotions differ between FWS?
2. Materials and Methods
To address the two research questions and to investigate if the effects are consistent across multiple studies and wine varieties, participants of one white wine and one red wine study were segmented using the fine wine instrument and their hedonic and emotional responses during actual wine consumption analysed. Study 1 [
41] investigated the effects of intrinsic and extrinsic product context factors on liking and emotional responses for the three FWS by evaluating three different white wine varieties presented together with three levels of product information. Study 2 [
47] investigated how intrinsic product and environmental context factor impact on the FWS’s hedonic and emotional responses when tasting Shiraz wines of differing quality levels in three separate consumption contexts (sensory laboratory, restaurant and at home). Furthermore, moods in anticipation of the wine tasting were assessed for the FWS.
Full materials and methods details can be found in [
41,
47] while all relevant details regarding the present research and data analyses are stated here.
All studies were performed in accordance with the ethical guidelines for scientific research at the University of Adelaide and approved by the human ethics committee (approval numbers: H-2013-048). Participants gave written informed consent prior to the tasting.
2.1. FWI and Emotion Scales
The 18 item Fine Wine Instrument Scale, comprised of 3 dimensions, connoisseur behaviour, wine knowledge and wine provenance, was used for consumer segmentation (
Appendix A Table A1) [
8]. This formed the basis for the Agglomerative Hierarchical Clustering (AHC) employed in the two studies described below. Emotions were measured on a 9-point Likert scale using AWEEL [
47]. To assess consumers’ wine liking a 9-point hedonic scale was used ranging from 1 = dislike extremely to 9 = like extremely.
2.2. Study 1 (White Wines)
2.2.1. Wine Samples
The aim was to select wines that were distinctly different from each other and cover different styles. Three commercially available Australian white wines, one Chardonnay (oaked), a Riesling and a Sauvignon Blanc, being the most commonly consumed white wine varieties in Australia, were selected by four wine experts (as defined by Parr et al. [
48]). Additionally, an unoaked Chardonnay was used as a warm-up sample to familiarise participants with the tasting procedure, as a control for first position order effects and to allow comparison between the two tasting sessions. Based on the results of a descriptive analysis [
49] of the wines, two wine descriptions were developed for each wine; a basic description, exclusively based on the objective sensory profile and a more elaborate description using more emotive sensory terms and including information about wine quality and background information of the winery (the full descriptions can be found in Danner et al. [
41]).
2.2.2. Tasting Procedure
In total, 126 regular white wine consumers (who consumed white wine at least once a month) evaluated the three wines under three context/information conditions, blind (no information provided), basic, and elaborate wine descriptions (for full descriptions see Danner et al. [
41]). Consumers completed an online recruitment questionnaire including basic demographic data, wine consumption behaviour and the FWI scale. To conceal the purpose of the experiments, the tasting was split across two sessions, at least one week apart, and consumers were informed that they were participating in two unrelated studies. In session 1, participants tasted the wines under blind conditions and in session 2, tasted wine together with the more detailed wine information. Then, 30 mL of samples were presented in coded ISO tasting glasses at 12–13 °C randomised within each session across participants. The sensory sessions were held in individual sensory booths, at the University of Adelaide, Waite Campus. Participants rated their liking and intensity of their emotional responses for each sample. The unoaked Chardonnay warm-up sample was presented first for both sessions. The statistical analyses did not show any significant differences between the two sessions in the evaluation of the warm-up sample indicating that wines were evaluated in a similar manner in both sessions.
Four participants failed to complete the Fine Wine Instrument questions; therefore, their data were excluded from analysis, resulting in a final number of participants of 122.
2.3. Study 2 (Shiraz Wines)
2.3.1. Wine Samples
One hundred commercially available Australian Shiraz wines were assessed by 8 expert judges (wine expert selection was based on the criteria outlined by Parr et al. [
48]) and placed consensually into one out of 4 quality categories—gold (an exceptional wine), silver (an extremely good wine), bronze (a good wine) medal and no medal but of sound quality (no faults present)—following the Australian wine show system [
50]. Based on the wine experts’ recommendations, sensory properties and availability, three wines from each of the four quality levels were selected for the consumer tasting.
2.3.2. Consumer Tasting, 3 Trials, 3 Different Contexts (Sensory Lab, Home and Restaurant)
A total of 349 regular Shiraz red wine consumers (who consumed Shiraz wine at least once a month) participated in one of three consumer trials. Each trial had a different set of four wines. Within each trial, the consumers tasted the same four wines, one of each quality level, under three contexts (sensory laboratory, restaurant, home). In the sensory laboratory, consumers evaluated wines in individual booths. Then, 30 mL of wine samples were presented in coded standard ISO glasses sequentially and monadically in randomised order. Water and unsalted crackers were provided for palate cleansing between samples. In the restaurant, participants were given 30 mL wine samples in coded ISO glasses in randomised order and instructed to evaluate all samples over the course of the dinner in that order. Consumers could bring a friend or family member to accompany them. For the home use test, the four wines in 750 mL wine bottles were given to the participants either after the previous tasting session or posted together with the home use questionnaire including detailed instructions. The wines for the home use tasting were masked in such a way that no identifying information about the wines was available to the respondents. Participants were instructed to taste all four samples on the same day with a main meal and to taste the wines in such a way as they would usually consume wine at home (e.g., use own glasses, preferred serving temperature, amount poured). The order of the tasting context was randomised across participants and they were not aware that they were tasting the same wines in the different contexts.
Additionally, to investigate differences in FWS mood in anticipation of the wine tastings, participants’ moods were captured using the Brief Mood Introspection Scale (BMIS) [
51] before tasting and evaluated liking, emotional responses and willingness to pay for the tasted wines. Three participants failed to complete the Fine Wine Instrument questions; therefore, their data were excluded from analysis, resulting in a final number of 346 participants.
2.4. Data Analyses
Agglomerative Hierarchical Clustering (AHC) using Squared Euclidean distance and Ward’s method was used to categorise consumers into one of the three identified fine wine segments (FWS) (No Frills, Aspirants and Wine Enthusiasts) based on their responses to the Fine Wine Instrument scale [
8]. Demographic data were analysed by Chi-square and z-test to investigate significant differences between FWS.
Repeated Measures Analysis of Variance (rANOVA), with wine sample and information condition as within-subject effects and FWS as between-subject factor, was used to analyse consumers’ liking and emotional responses in Study 1. Previously, Danner et al. [
47] showed that the inclusion of a trial as a random effect does not improve the ANOVA model, so a similar rANOVA approach with wine quality and context as within-subject effects and FWS as between-subject factor was chosen to analyse liking and emotion ratings of Study 2.
All statistical analyses were performed with SPSS Statistics (Version 25, IBM Corporation, New York, NY, USA) at 5% level of significance.
4. General Discussion and Conclusions
This is one of the first studies investigating wine consumer lifestyle segmentation while capturing participants’ responses when actually tasting wines, as opposed purely to online self-reported surveys. Across the two studies, it was found that tasting or consuming wine elicits different hedonic (research question 1) and emotional responses (research question 2) in the three identified consumer segments derived from the fine wine instrument (FWI).
The identified Wine Enthusiast segment generally liked the tasted wines more and stated more intense positive emotions when tasting or consuming wine compared to the No Frills segment, with the Aspirants ranging in between. Furthermore, Wine Enthusiasts were more discriminant regarding wines preferred and were more emotional about higher quality wines (Study 1) and more complex wines (aged and oaked Chardonnay, Study 2). Therefore, if wine quality and complexity are of interest to wine producers, it might be beneficial to focus research efforts on Wine Enthusiasts. In summary, these results indicate that “domain-specific” lifestyle segmentation, in this case the FWI, is a reliable segmentation method resulting in meaningful segments.
Interestingly, although significant main effects of context were observed for both studies, these effects were consistent for all three consumer segments, with only minor differences in the emotional responses of adventurous and enthusiastic, indicating that FWS are similarly influenced by the tested contexts and information conditions. A possible explanation could be that consumers were only segmented on their responses to the fine wine instrument. If another segmentation base encompassing other lifestyle components was chosen, the impact of context and information conditions might differ.
Consistently across both studies, negative emotions were only elicited to a very small extent, and no significant differences were observed between FWS. This is in agreement with previous studies showing that commercial products of sound quality are generally associated with moderately intense positive emotions, but related very little with negative emotions [
59,
60]. Only if faults or less desirable sensory properties are exaggerated, do negative emotions increase in intensity and discriminate between samples, e.g., intense astringency [
57] or green aroma [
45]. Although the quality of Shiraz wines in Study 2 varied, even the “No medal” wines were of sound quality and no faults or off-flavours were present.
It must be noted that data were collected exclusively in Australia and, although segmentation outcomes show considerable overlap with other markets (e.g., Spain [
46] or South Africa [
26]), further studies are required to confirm that the results are transferable to other countries and markets.
Recently, Mora et al. [
58] segmented consumers based on their personality using the Big Five Inventory questionnaire (BFI) [
61] and found differences in hedonic and emotional responses between the personality segments when tasting a wide range of different wines. Investigating the personality traits of the identified FWS could be of high interest for future research and might assist in gaining a better understanding of how consumers’ wine knowledge and involvement influence purchase decisions.
Wine involvement has often been associated with wine knowledge. The current studies, in agreement with Calvo-Porral et al. [
46], also show a significant emotional aspect to wine consumers’ involvement. Therefore, creating positive emotional experiences around wine might be a promising opportunity to increase consumers’ wine involvement.