Abstract
International marketing research has always been considered more complicated than domestic marketing research, as the collection of accurate and reliable data in international markets is an even more intricate and complex endeavor (Craig and Douglas 2005). Differences in language, culture, and stage of cultural development, on the one hand, and the chance of selection errors due to lack of sufficient and reliable information, on the other, make the use of standardized research approaches and methods difficult and often infeasible. In a complex global world, the call for more rigorous methods to analyze and interpret international data is nevertheless more necessary than ever. The four articles in this special issue address different gaps in the current toolbox available to international marketing researchers. Before elaborating on these four studies, I would first like to discuss a couple of evolutions in international marketing that make rigorous research in global marketing even more challenging and exciting now and in years to come.
The explosive growth of world trade has unleashed an increased need for information about markets throughout the world. Interestingly, this need for information is typically examined from the perspective of large, corporate multinational enterprises. Nevertheless, the rise of global value chains and the digital transformation of the economy offer new opportunities to participate in the global economy for small and medium enterprises (SMEs), which until recently were underrepresented in international trade (Organisation for Economic Co-operation and Development 2021). The improved access to digital technologies substantially lowered SMEs’ barriers of entry into global markets. This evolution allows SMEs to internationalize at a fraction of the cost, making it easier to find customers abroad and take part in global trade and export. Paradoxically, in spite of this digital revolution enabling firms to find new export opportunities, collecting information about and benchmarking performance in foreign markets remains challenging for most SMEs. The lack of dedicated workforce and limited resources on the SME side typically reduce the extent of both external and internal knowledge that SMEs can rely on to assess and adjust their foreign operations. This is all the more difficult because, despite prior research on the measurement of export performance, no specific measurement frameworks have been attained. Moreover, both financial and nonfinancial elements play an important role, making it even harder to construct robust holistic metrics. Because smaller firms also face greater challenges in navigating foreign markets, getting better tools to track their performance in these markets is paramount in helping SMEs—and the many people they employ—take advantage of these opportunities in international markets and succeed in the long run.
In addition to these challenges, companies face increasing cultural diversity of marketing operations as they expand abroad, making it all the more important for researchers to collect information on the changing lifestyle and consumption patterns across the world. Although global accessibility of information has improved substantively over the last decade, several issues remain. For one, with an ever-expanding pool of countries for which data become available, the question arises whether classic metrics need to be updated and revised. Given the increasing availability of information from typically underresearched regions and countries, the heterogeneity in the measures and metrics using this expanded pool of information requires further examination. For example, what are the implications for the measurement of cultural values and cultural distance, some of the most-used concepts in international marketing (Steenkamp 2001)? Traditionally, notions of culture are defined by geographical country. As these nation boundaries blur and change, deterritorialization of culture may follow (Hernàndez i Martí 2006). This raises the following questions: To what extent do cultural value and distance metrics based on the Hofstede (1980) or Schwarz (1999) dimensions need to be revised? Moreover, how should we interpret changes in cultural values over time based on these Hofstede or Schwarz metrics? Do changes in these metrics truly reflect evolutions in cultural values, or are these evolutions an artefact of the increased heterogeneity in the set of countries and regions sampled? Moreover, are nations still the relevant entity to capture culture? Assessing the sensitivity of the extant body in international marketing to these evolutions is most definitely required.
Furthermore, an increasingly globalized world implies that our knowledge of international marketing phenomena can no longer be derived uniquely from research and data conducted in high-income, industrialized countries. This implies not only that more research is called for in emerging markets but also that existing measures and methods to collect data in these markets need to be updated and more stringent corrections in data analysis are required. Compared with developed markets, data collection in emerging markets is more challenging, and the collection method itself requires careful consideration. Even for secondary data, as provided by the World Bank, recency and accuracy biases may arise, calling for care and corrections when using these data (Burgess and Steenkamp 2006). In particular, primary survey data—often the traditional workhorse of data collection in international markets—can be marred by the absence of uniform sampling frames. Moreover, due to the extreme heterogeneity within these markets, the potential for severe selection biases can become even more striking, as the data may not represent the vast majority of the population. This issue may be even more prominent in retail and household scanner data that reflect the shopping behavior of only the (rising) middle classes (Burgess and Steenkamp 2006). To provide insight into the behavior of the vast majority of the market, beyond the middle class, other methods of collecting shopper and purchase data are required.
Several calls for an extended body of work on the bottom-of-the-pyramid market have been made and followed to delve more deeply into this within-country heterogeneity (see, e.g., Burgess and Nyajeka 2006; Chandy and Narasimhan 2015, Von Janda, Shainesh, and Hillebrand 2021). The bottom of the pyramid represents approximately four billion people earning an income of less than U.S.$2 per day and consists of diverse segments with different decision-making processes, likes, and needs (Prahalad 2012). Studying the bottom of the pyramid raises a multitude of questions on how existing methods and data collection should be adjusted to accurately capture insight into attitudes and behaviors. Even when adequate data are available and accessible, they are frequently of poor quality, relying on surveys that require total comprehension of the underlying construct on the part of the respondent, which may be lacking among poorly educated respondents (Grosh and Glewwe 2000.) As such, traditional Likert-scale based data are difficult to obtain; even when such data are collected, a lack of understanding of the concepts and constructs can lead to severely skewed data. This reduction in variance makes the data useless for interpretation and creates situations that are ripe for response bias (Chandy, Hassan, and Mukherji 2017).
The use of less structured data collection methods through in-depth interviews and video observation may therefore be necessary (Penz and Ghauri 2005). Whereas classic qualitative data collection techniques have been applied to analyze data in international and emerging markets, machine learning techniques using text and video as input may provide new opportunities to complement and analyze these data. Ultimately, this may allow for new insights as well as expand the scale and scope of data collection opportunities for this large, often neglected segment. Interestingly, in emerging markets, including some of the poorest countries, governments and companies have used artificial intelligence and machine learning more extensively than in so-called developed markets. For example, machine learning algorithms are frequently used to predict the probability of default of potential borrowers and to solve critical development challenges such as the provision of financial services to underserved populations (Strusani and Houngbonon 2019). Nevertheless, extant research in international marketing has largely ignored the wealth of big data that can be obtained in emerging markets (Burgess and Steenkamp 2006).
The widespread use of mobile devices by the bottom-of-the-pyramid segment opens up interesting avenues for future research in international marketing. Mobile apps can be used to identify and communicate issues and needs more easily by, for example, uploading pictures rather than by using extensive text, and they also provide a rich source of data. In addition, this method of data collection may reduce social desirability biases. In classic data collection settings, respondents in emerging markets are typically surrounded by family and neighbors, making these respondents less inclined to respond to sensitive questions. Big data collected through mobile devices, however, may reduce these types of influences. Interestingly, emerging markets may even stand to gain more than developed markets from the potential offered by machine learning and big data analysis, as these markets have less legacy infrastructure, which may make it easier to leapfrog traditional technologies in virtually all sectors of the economy. Because these markets have historically suffered from data poverty rather than an abundance of data, the intensity of need—for the largest numbers of people—is possibly the greatest in these contexts (Chandy, Hassan, and Mukherji 2017).
Still, advances in technology simultaneously facilitate and complicate collection of data on a global basis (Craig and Douglas 2005). Even with more data available on a wider set of countries and segments, obtained through multiple sources, the issue of comparability across markets and segments remains one of the most challenging for international marketing researchers. Knowing how to address and correct the potential for a variety of biases thus becomes all the more important.
The four articles selected for this special issue broadly examine how methods and metrics in international marketing research can be updated and modified in light of these new challenges and opportunities in the global market space.
First, Baumgartner and Weijters (2021) reflect on the common method variance bias in international marketing research. To do so, they first look into whether the extant literature in international marketing has adequately addressed the issue. Second, they examine whether alternative corrections substantially alter substantive findings. Common method variance is always a concern, regardless of whether a study is embedded in an international context. As soon as several metrics are obtained from the same respondent in a survey, the respondent’s response style across all questions can make it difficult to distinguish between dependent and independent variables, leading to erroneous conclusions. In an international context, this risk may be even more extensive, as respondents from different countries are ingrained in distinct cultures, and their response patterns and expressions of (dis)agreement on specific concepts and constructs vary significantly. Thus, people’s response style to the survey can be confused given the substantive differences across different countries. By taking stock of current practices in the Journal of International Marketing, Baumgartner and Weijters conclude that researchers currently overly rely on marker variables and the Harman one-factor test to detect common method variance, leading to the inaccurate conclusion that common method variance is not an issue. When working with data collected in emerging economies or derived from the bottom-of-the-pyramid segment, these issues may become even more prevalent. The authors therefore propose a post hoc procedure for incorporating method effects in international marketing research using a directly measured method effect or an inferred method factor as a control variable at the item level. When using a latent method, the researchers recommend the use of a research design that makes the specification of a method factor meaningful. In future research, it would be interesting to see how and to what extent common method biases exist outside the scope of primary data research. Is common method variance more or less present when using different big-data type sources, and how can we translate detection and correction methods to this setting?
Two further articles in this issue focus on the cultural distance construct. As discussed previously, the growing heterogeneity in countries sampled can have implications for the observed cultural values and cultural distance metrics. Messner (2021) explicitly considers the implication of this growing heterogeneity. In the tradition of Kogut and Singh (1988), the heterogeneity between the set of sampled countries is corrected for by scaling each cultural value by the variance observed in the sample of countries. However, with growing heterogeneity, the difference in cultural values between the same set of countries may change over time simply because the sample on which time-specific observations relied became larger and more diverse. Invalid conclusions can therefore be drawn about how countries diverge or converge on culture over time. With a strong need for more longitudinal work in international and cross-cultural research, measurement tools of cultural values and distance are required that do not confuse evolutions in sampling with cultural differences. To that extent, Messner introduces a geometrical difference measurement of cultural distance using the angle of heterogeneity. The metric does not necessitate independence of the cultural value dimensions, nor does it require the dimensions to have equivalent measurement scales. Most importantly, it produces identical results independent of the number of countries the dimensions are based on, making the metric more appropriate for longitudinal studies.
Also focusing on cultural distance, Griffith, Dean, and Hoppner (2021) further delve into the idiosyncrasies of various operationalizations of the cultural distance construct. The authors demonstrate that differences in operationalization of cultural distance can lead to substantive differences in the observed effects of cultural distance. The justification of using the right operationalization given the research question and context therefore becomes all the more important. Not only does the selection of the entity, individual or aggregate, play a role, but it is key to carefully consider which dimensions of cultural distance substantively play a role. Moreover, like Messner (2021), Griffith, Dean, and Hoppner caution against overly relying on the Kogut and Singh (1988) operationalization and urge researchers to specify which operationalization was used. Ultimately, better insight into the fit between the metric and the substantive issue should help researchers avoid choosing an operationalization simply based on the “best” p-value. This is a useful notion for both researchers and future reviewers of international marketing manuscripts. Ultimately, the most appropriate operationalization, rather than the best p-value, will help advance the body of work in international marketing.
Finally, Sadeghi, Rose, and Madsen (2021) turn to the needs of small and medium-sized enterprises (SMEs) in international markets. As SMEs’ export performance reflects both financial and nonfinancial goals, it becomes difficult for them to measure their export performance consistently across markets and over time. The authors therefore propose a customized measure of perceived export performance. Their framework considers managers’ specific priorities through explicit incorporation of manager- and firm-specific differences in the types and importance of goals, indicators, and benchmarks. As such, a more holistic measurement approach is provided that is tailored to individual firms and reflects firm-specific idiosyncrasies. In adopting this approach, firms develop a framework that uses their employees’ assessments in a structured way that allows for comparisons between markets and over time. As such, using their approach, SMEs will be able to pinpoint more quickly and accurately both export opportunities and challenges and redeploy their limited resources more appropriately to markets that offer more long-term potential. Over time, this approach will also allow the firm to create a set of internal benchmarks to assess its export performance, even for markets for which little to no external knowledge is available. In the end, it will allow SMEs to operate more nimbly in an ever-growing global market.
In all, the four articles in this special issue offer useful and practical insights for both academics and practitioners in international research. Given the ongoing growing scope of countries and businesses operating in the global realm, the need for more accurate and robust research methods and tools that allow for inferences across markets and over time is increasingly prevalent. These articles demonstrate that the correct selection of tools is vital, as results can be affected substantively. With the torrent of new data (text and visual), collection tools (e.g., apps) and analysis methods offered by machine learning coming into the market, it is all the more important to understand how the selection of method and metrics impacts and advances the body of knowledge in international marketing. This special issue offers researchers a beacon regarding how to proceed and, at the same time, calls for more research as data sources become larger and broader.