The Bigger They Are, The Harder They Fall: Pros and Pitfalls of Multi-Country Research

I’ve been heavily involved in international marketing research since the ’90s. Here are some of the things I’ve learned along the way. First, I should make clear that this brief article is aimed at providing practical advice to marketing researchers – I am not writing to scholars or researchers working for the United Nations.

Secondly, it is focused on primary research and quantitative research, which are my specializations, and written from the point-of-view of an external consultant. Lastly, for readability, I am using “global” when regional in some instances would be more precise.

“Globalization” was the “AI” of the 1990s – everyone talked about it, but few really understood what it was. The ’90s saw the emergence of huge multinational marketing research companies, most of which in one form or another are still in existence. We’ve moved on from globalization, in the sense it doesn’t generate much buzz anymore, to Big Data, IoT and, in its latest incarnation…AI. Gigantic continuous tracking studies conducted in dozens of countries concurrently seem to have fallen out of favor for various reasons – cost being one – though they are not entirely a thing of the past.

Why conduct multinational marketing research? At one point, many multinationals were absorbed with developing global brands that, with a bit of glocalization tweaking, were essentially the same product with the same positioning everywhere. The rise of powerful local and regional brands not headquartered in Western countries or Japan has threatened this strategy, which I always thought was more fad than sound marketing. Tastes do vary, after all. However, it does make sense to keep tabs on vital marketing statistics such as spontaneous brand awareness that are difficult or impossible to obtain through secondary or other “Big Data.” 

Analysis multi-country data is usually descriptive, as is true of most market surveys. However, perceptual mapping, conjoint, key driver analysis and segmentation are also conducted, especially the last of these. I should add that, even with all the AI hype we must endure these days, many marketing researchers and decision makers at clients are largely unfamiliar with multivariate analysis, and that the level of research sophistication in general is uneven around the world.

Here are some broad guidelines I find useful in international studies. 

Make certain you understand the objectives of the research, who will use the results and for what purposes they will be used. Tensions between global and local country management teams are almost a given. What will be of use to one is very often of little or no use to the other. It is not unusual for local companies to be forced to fund research in their country that they will never use or, worse yet, implement policies based on it they strongly disagree with. More than once I have been asked to reanalyze local country data for the local country team. All this said, do not allow yourself to be caught in the political crossfire. I have…and it hurts.

Make sure your research team is ethnically diverse and whenever possible have people “on the ground” on the local countries. The world is all one place…except it isn’t. Be very mindful that someone can have an MBA from Harvard, and look and dress much as you do yet be very different in important ways. Though middle-class people in many countries are superficially Westernized, do not assume that it’s OK to play by Western (or your own local) rules. Do not assume, because someone is fluent in your language, that they truly understand or identify with your country or national culture. 

It is important that, whenever possible, team members include veteran researcherswith experience in international research. It is a mistake to “dump” multi-country studies on junior staff.

While this may seem self-evident, make sure you are clear who the contact person or persons within the client organization are and who in your company should be copied on correspondence. Getting lines of communication straight is very important but not always obvious, especially in multinational studies. Make sure everyone has the time differences right. It’s very easy to check this and public holidays on the internet. Speak slowly and write clearly. Avoid slang and jargon.

Have questionnaires professionally translated and back translated. Brand names can be easy to mix up, but an even bigger sand trap are psychographic items. These are often very difficult to translate and may make little sense to non-Westerners, or be misinterpreted by them. Some items may also be considered offensive or too private to answer.

Questionnaires in multi-country studies can be very long since certain questions may have relevance only to some countries, or groups of countries, due to cultural similarities or because of management structure – countries may be grouped by the client into sub-regions, for example, and brands managed at that level. This means only parts of the questionnaire need to be translated into all local languages. It also complicates reporting. As you might imagine, the administrative burden of multinational projects can be considerable. This is true even of trackers since at least slight changes are typically made from wave-to-wave.

Scale usage and response styles in general can vary substantially by ethnicity, even within a single nation. I’ve listed many books and other resources in my library, and the sections Sampling and Survey Research and Structural Equation Modeling and PLS are probably most germane to our topic. (Trade-off methods such as conjoint MaxDiff are not a panacea.)

Quality online panels are not ubiquitous and face-face interviews are still conducted in many countries. Sample and questionnaire design must account for these inter-country differences. Be aware of inconsistency in fieldwork quality. Even the biggest networks have patchy fieldwork capabilities, and in some markets, outsource all local work.

Weighting is a sticky wicket. If you don’t assign country weights – that is, by default assign each country a weight of 1.0 – the impact of small countries may be higher than their business significance. But what to weight by and what is big and what is small? A country can have a massive population but the size of the category being researched in that nation may be quite small, and vice versa.

Demographics such as occupation, income and socio-economic class must be adjusted to facilitate meaningful comparisons. Sometimes it helps to have two sets of tabulations of these demographic variables, one coded for the global team and one for the local team.

In many ways, the world is a smaller place than it was two decades ago, in part because of the Worldwide Web but also because of migration patterns – people from any place on the globe are now living any place on the globe. Legal and Financial integration of multi-national corporations is not the focal challenge it once was, but cultural integration of management teams remains an obstacle and will until we really are all the same (a day I do not welcome). 

The Journal of International Marketing is a good source for the latest scholarship on international marketing. This short article of mine, Branding Asia, may also be of interest and the book, Hard-to-Survey Populations, which I review here is pertinent to multi-country research and a joy to read if you’re a researcher at heart.

MR Realities is series of audio podcasts in which Dave McCaughan and I discuss a wide range of topics important to marketing researchers with special guests. One of these topics has been international research, and here are the links to those podcasts.

This has just been a snapshot of a big topic, but I hope you’ve found it interesting and helpful!

Arrange a Conversation