Behind the scenes at the world’s largest wine consumer survey
11 years ago this month, in 2007, just as the financial crisis was brewing up in the world’s largest economies, Wine Intelligence launched the first multi-market omnibus consumer survey dedicated to wine drinkers. By this point we as a company had run quantitative consumer research, in several markets, and for several years, but this was our first go at something co-ordinated, and large scale omnibus launch. Crucially, it followed a significant investment by us in establishing exactly what kind of people we should be talking to in each of the 10 markets we surveyed in that first wave.
This is what we mean when we say our data is ‘representative’. This word is easily thrown around in the world of market research, partly because it is vital to the correctness and credibility of this kind of data. Put simply, collecting a sample of, say, 1,000 interviews in a market and then claiming that the views and behaviour of those 1,000 ‘represent’ that of the market as a whole is only possible if you have a reasonable idea of the profile of the people who should be in the survey from the overall population of interest, and then make sure that those characteristics – age, gender, income level, wine drinking behaviour – are matched in the sample you take. Following the tenets of statistical theory, such a sample (known as a quota sample) then will give you a better picture of the population of interest than one where you just took the first 1,000 answers you got, irrespective of who these respondents were.
Of course there are problems with quota sampling too. For instance, how do you get the quota in the first place? And how do you know it’s correct? Nowadays, Wine Intelligence runs Vinitrac® surveys in 32 countries (latest addition: Paraguay) and part of the challenge we face every year is keeping our quotas up to date – or even, in the case of new countries, establishing them in the first place. In order to make sure all surveys are representative, extensive market research has to be conducted before we even distribute the surveys. In each market we need to run calibration studies. We tend to use very well established and far-reaching general omnibus studies (an omnibus, like Vinitrac® itself, is a survey where anyone can buy a space and ask questions of their own choosing). The data coming out of an all-adults survey with national representative reach will tell us, with a good degree of accuracy, the information we need to set our all-important quotas, which determines how many people in each age and gender category, for example, we need to target in order to be nationally representative.
We also use calibration studies as one (though not the only) source of data to tell us whether the wine drinking population in a given country might be increasing or decreasing. The percentage incidence rate (% IR) of wine drinking in the calibration study does fluctuate of its own accord, so we need to triangulate any finding here with corresponding data points, such as the country’s own published statistics on wine drinking, or World Health Organisation data, and correlating measures such as per-capita consumption and wine volumes sold, as measured by our partner agency the IWSR.
IRs are given as a percentage, and the calibration process is altered depending on the type of market. ‘Mature’ markets are those in which wine has reached its potential with stable or declining volumes, whilst ‘established’ markets tend to have a strong historical growth that is now tailing off. For these two classifications, the IR represents those in the total population who drink wine on a regular basis and who are above the legal drinking age, in all regions and cities.  (For more information on market classification please see Global Compass.)
Top 5 Vinitrac® wine markets with the highest incidence rates of regular wine drinkers amongst the adult population:
 Vinitrac map
Size of pie = Total population
Blue portion of pie = Incidence rate (Proportion of total population who are Regular Wine Drinkers)
Number in pie = Regular Wine Drinkers in millions (i.e. drink wine at least once per month)
Source: Wine Intelligence, Vinitrac® Calibration Studies
By analysing the demographics of this segment of the population, we are able to establish our quotas so that our survey sample reflects the gender and ages of these regular wine drinkers.
Demographic profiles of the top 5 Vinitrac® mature / established markets with the highest incidence rates of regular wine drinkers:
 Vinitrac chart
 Source: Wine Intelligence, Vinitrac® Calibration Studies
Vinitrac® remains the largest ongoing online study of wine drinkers in the world, and since 2013 we have collected responses of approximately 500,000 wine drinkers. It is not only an incredibly powerful tool, but also one that allows us to be much more confident in the reliability and accuracy of our surveys due to Vinitrac® samples being nationally representative of each country.
Extraordinarily, some well known studies in the wine category still don’t bother with this step. So if you are being presented with survey evidence, and want to test its validity, ask about how it deals with the question of representativeness. The person presenting it is either going to be delighted at being able to give a technical answer (you have been warned); or shift uneasily and stare at the floor. Either way, you will have your answer.
While the calibration studies are very useful to us in building a better and more accurate survey, they might also be useful as a business tool. If you would like to find out more about your wine consumers on a national level, or want to make comparisons across several markets, our calibration and quota sampling process is what makes this possible. For instance, you may be interested in a specific age group, or maybe want to focus on how your brand is received in a particular city.
Wine Intelligence will be launching the next wave of Vinitrac® this coming March, so please submit your question requests before Friday 23rd February 2018.
Future 2018 Vinitrac® waves include Sparkling Wine in May and Still Wine in July and October.
Author: Jasmin Blockey