In order to stay ahead of the competition in a constantly changing global environment, clients demand information that is timely, reliable and accurate.
In the case of a global market research study, the cost of being timely (i.e. fast) can be a lack of understanding of the impact of cultural factors on the data – its reliability and validity. Cultural impact is perhaps best understood using an “emic” approach that believes “attitudes and behaviour are unique to a culture and best understood in their own terms” (Craig, Douglas, 2005).
The emic approach, however, is somewhat time consuming and less practical in today’s fast-paced environment. This often leads the international marketers to prefer the alternative “etic” approach. In this school of thought, the researcher attempts to measure universal attitudes and behaviours using culture-free measures. In practice, this is a research design developed in one culture and then applied to foreign cultures with minimal or no adaptation to local culture. This approach brings with it the speed advantages as well as “at face value” comparability. Potentially, however, it raises a number of issues related to the relevancy of certain measures within any one culture.
Ideally, a culturally balanced study ought to be preferred to ensure there is no bias in terms of any one researcher’s (or client’s) culture, while maintaining the original goal of conducting a global study looking for similarities. The question then is: to what extent should we adapt or localise that which is unique to a given country? And how far should we emphasise the common elements across countries? In general, there are three areas that need attention for localisation: demographics, behaviour and market dynamics, and attitudes and psychographics (Craig, Douglas, 2005).
On the face of it, demographic characteristics ought to be less problematic, as they are often universal, such as gender and age. However, the way the questions are posed may differ dramatically. A global study conducted by SSI found that asking for age in different formats gives different results in China as seen in the table below. Relative to the answers given in open numeric format, the difference was smaller when asking for the date of birth in the format commonly used in China (YYYY-MM-DD):
Asking for other background information such as income, education, or occupation brings more challenges. For example, people in some countries, such as Indonesia, would avoid questions about their income for either personal reasons or tax purposes. In this instance, asking their household expenditures may give a more accurate answer. Even if income itself is the right measure, questions arise: how should the time-frame be defined (weekly, monthly or yearly)? Should we measure take-home pay or gross income? What elements of income should be included? In some countries, bonuses may be given at the end of year that may not be counted as monthly income.
In some countries, alternative solutions measuring purchasing power instead of income are employed. These use social grade or economic classification (SEC). For example, a European Social Grade developed by ESOMAR is applied in some countries in Europe. This grading system takes account of the occupation and age of terminal education of the main income earner. Similarly, market research organisations in Latin America have created an SEC based on education and occupation of the main income earner in the household. The Market Research Society of India (MRSI) developed a grading system based on the education of the chief earner and ownership of a number of consumer durables. Indonesia has a social class system based on monthly household expenditure. The Vietnam and Thailand Market Research Societies developed their social class based on family or household income. One of the biggest challenges is maintaining and monitoring the social grade system in order to be still reliable and relevant, especially in rapid-growth, developing countries.
Asking for education levels is also not always comparable from one country to another, because the schooling system is different across countries. Practical alternative solutions exist; for example, asking for the number of years in formal schooling or the age at which full-time education was completed. However, this is not easy to either compare or to maintain globally, as governments change education systems to meet local requirements.
Similarly, asking for specific occupation is complicated due to the difference in government systems and bureaucracy, such as solicitors in the UK and Australia, but lawyers in the US and Singapore. Perhaps asking for broad categories of occupation gives more comparable segments; for example, self-employed, professional, white-collar workers, blue-collar workers, and senior management.
As a final example, household size is another variable that has to be understood across cultures. For example, extended family structures in Asian countries such as India, China, and Indonesia include multiple primary units of husband, wife, and children, as well as grandparents, other siblings, cousins, unmarried aunts, uncles, and so on. This implies that the household size may not be comparable, and that the results need to be read cautiously.
Behavioural and Product Market-related data
The need to localise consumption and purchase behaviour questions is heavily related to the economic environment of the country. For example, in developing countries where the traditional retail trade (i.e. small shops or wet markets) dominate, consumers may rely on the shop owners’ advice on products or brands to purchase. On the other hand, consumers in developed countries may tend to be price conscious due to price competition among supermarkets, hypermarkets, or discount stores. This suggests that localisation is needed to tailor the specific retail infrastructure in both economic environments, and that care must be taken in interpretation of the results.
As another example, having a refrigerator allows consumers to stock more fresh food at home and hence reduces their frequency of shopping. This may not be the case in developing countries where doing grocery shopping on a daily basis is not uncommon. Not asking about refrigerator ownership (since it is universal in the researcher’s own culture) may lead to non-comprehension of the results.
In many Latin American and Asian countries, domestic helpers and/or cooks are common among middle- and upper-class families. They often know better the usage of the cleaning products or kitchen appliances, and therefore their opinions matter and influence the products or brands to purchase. And yet they are not chosen to be the respondents, or may answer differently for their own personal household.
As another example, the practice of gift-giving plays a key role in maintaining the social structure in Japan. Gifts are carefully selected and often matched to the status of recipient, and the wrapping is equally important. Therefore, attributes related to products or brands that are positioned as gifts need to be customised to reflect this local practice.
Another concern related to consumption and purchase behaviour is the product definition or description that may be defined differently in some countries. This becomes yet more challenging given product life cycles and the unclear category boundaries. For example, the soft drinks category may be classified differently across countries. Initially, it was dominated by colas and sodas; however, it has expanded. In the US, energy drinks are considered as soft drinks, whilst they may not (yet) be considered as soft drinks in Asia. On the other hand, flavoured soybean milk is considered a soft drink and is widely available in Southeast Asian countries. This product may be unavailable or have limited distribution in other regions, or may not be considered as a soft drink. Take also the example of fresh fruit juice. Are these considered to be soft drinks? They may be in some markets. Using category names is fraught with difficulties in international research; adding examples of brands within the category only adds to the problem. Not understanding the local nuance means not understanding the data.
Where the category is essentially self-defined, further problems occur. “Premium skincare products” is one such example; one person’s “premium” is another person’s “everyday”, and vice versa. Another study conducted by SSI saw a majority of Chinese women claiming to purchase “premium” skincare products. Examination of the brands mentioned revealed that only 16% of the answers would be considered “premium” by the manufacturer, coming of course from another culture.
Attitudinal, Psychographic, and Lifestyle data
In multi-country research, designing questions related to attitudes, psychographics, and lifestyles is one of the most challenging problems for the researcher working alone in his or her own culture. Some concepts may be unique to a given culture but irrelevant to the others. For example “giri”, meaning obedience and respect, is a value unique to a Japanese environment. Excluding this value from the attributes list in Japan may cause problems in explaining inconsistent or unexpected findings. Including this attribute in a survey in the UK or US may equally lead to incomprehensible results.
Other concepts may nominally exist across countries but be expressed (and therefore understood or associated) in different ways. Craig & Douglas (2005) use the example of “aggressiveness”. This may be expressed in terms of physical behaviour (i.e. fighting), but it may, in a different culture, be expressed in terms of verbal behaviour such as shouting.
In marketing, too, we see this differential perception. Take the example of “value for money”. In the originating culture, this is meant to describe better value than what one is paying for. In the receiving cultures, it may be perceived as a cheap product. In such instances, perhaps an explanation or different phrase, rather than a direct translation, is needed in order to have a comparable result.
These differences in culture can run much deeper than mere semantics and be more pervasive in their impact. Take, for example, understanding the relative importance of self and others. This is different in the extreme between Asian and Western societies (Markus & Kitayama, 1991). In Western culture, the individual or self is perceived as independent and self-contained, and as having abilities, motives, and values. In many Asian cultures, such attributes may be perceived less positively, as the norm centres around fitting in and a harmonious interdependence with others. Products with branding that emphasise “independence” may need a delicate approach in communicating the products, or they may risk a failure, not because of a problem with the product, but with the association.
Many background and demographic characteristics are relatively simple to adjust in order to generate comparable data across countries. However, for many purchase and consumption measurements it might not be possible to have strict “apples-to-apples” comparisons, due to structural variations from one country to another. Some of these variations can be researched independently through desk research, and the questions or answers then adapted. The greatest challenge is likely to occur when designing questions or attributes related to attitudinal and psychographic measures. Local advice can be invaluable in this regard, but it must be advice based on equality and a shared common knowledge of the objectives of the research. Otherwise, it runs the risk of merely being an aid in translation that leads to misinterpretation.
By Ati Sinaga, SSI Global Knowledge Group
First published in Asia Research Magazine, Q2 2014