Practical Winery
65 Mitchell Blvd, San Rafael, CA 94903
phone: 415-453-9700 ext 102
1 · 2 · 3
Summer 2011
The few studies conducted have shown that more than 50% of consumers read back labels and find them important in their selection.1 Existing research on back label content is limited mainly to small studies of narrow sets of attributes.
Ten years ago, the Australia & New Zealand Wine Industry Journal published a study showing that different back label statements created different perceptions of a wine’s quality.5 Beyond these, the other studies only investigated one or two attributes, usually in the context of front and back labels. None of these studies were able to quantify the importance of different types of back label statements or test them against a consumer’s actual wine choice, especially with different prices.
We developed an experiment to test 10 different back label statements in an incomplete block design along with four different prices. A sample of 331 wine consumers was recruited in May 2008 for a
wine tasting research project in North Sydney, and were given the back label experiment as a part of this project. The sociodemographics of the sample were similar to those of the general Australian wine consumer population, when compared with the Roy Morgan single-source data of over 50,000 Australian consumers.
The respondents were given a range of socio-demographic and wine consumption questions along with the experiment. They were asked to consider purchasing a wine for a special occasion, which represents approximately 25% of the purchase occasions in Australia,2 and given a printed page with four different back labels and prices for a Shiraz wine.
The statements were taken from previous research and from common back label text. We added one on ingredients, since this is under consideration in both Australia and the U.S. Each back label had between 2 and 10 statements. Respondents chose the wine they would be most likely to buy from 16 different sets of four labels, which allowed us to compute the effect of each of the statements and price on the probability of choice. Respondents were asked, “Would you really purchase your choice?” to test the realism of the experiment. Eighty-seven percent said they would buy their choice.
A latent class choice model was used to simultaneously create clusters and utilities for each statement in each cluster. The best solution had the 331 respondents grouped into five different clusters, with strong differences in the importance of each of the statements and price across the clusters (Table II). The final column in Table II shows the average value for the sample.
Overall, price accounts for 66% of the importance, with the back label attributes accounting for the remaining 34%. Listing the “ingredients” had the highest value — though
with a negative influence on predicted choice — followed by history, food pairing, and the elaborate (or longer version) description of the wine’s taste.
The clusters differ substantially in the importance of price and back label statements. Ordered from 1 to 5, the clusters with lower numbers preferred lower prices and those with higher numbers preferred medium and higher prices.
In two clusters, price is dominant (C1 — low price and C3 — medium price). In other clusters, the back label statements have a larger effect. Perhaps the most interesting is C5, where ingredient labeling has a very large negative effect on the probability of purchase. The advent of ingredient labeling would certainly affect people in this cluster.
The lowest price C1 cluster (which is about 31% of the sample), values food pairing and not much else. The C2 cluster prefers wines at either $13.99 or $19.99 and strongly values information on winery history, production methods, and both simple and elaborate taste descriptions. C3 prefers prices around $19.99 and is influenced somewhat by simple taste descriptions and food pairing. Consumers in C4 choose prices around $25.99 for a gift or special occasion and value elaborate taste descriptions and history, but are negatively influenced by ingredient labeling.
Overall, those clusters for which price is the predominant and almost only choice-driver represent about half of the population (C1 and C3). About one-third of frequent Australian wine consumers (C2 and C4) can be positively influenced by back label information, especially history and elaborate taste descriptions. About one-third of frequent Australian wine consumers are adversely affected by stating ingredients on the back label, with a small share (about 13%) of consumers (C5) refusing to choose labels including them.
Previous · Top · Next