Marketing research education in the Big Data era

Redefining the content of quantitative marketing courses

Next to this name-based repositioning, the content of analytics courses in the marketing discipline should include transactional data, which have become of increasing strategic relevance. Furthermore, Big Data, derived from, for example, online interactions with customers, are also used by various organizations. Note that transactional data cannot be considered as Big Data. In terms of the well-known three V’s, the volume of transactional data is not particularly large (gigabytes and sometimes terabytes, but not petabytes), there is limited variety as transactional data are mostly structured, and velocity does not apply as transactional data are usually processed in batches at set time intervals.

Particularly relevant for the current note is that most marketing research courses do not teach students extensively about transactional data or Big Data, but emphasize survey data (Nunan & Di Domenico, 2018). At the same time, computer science departments are rapidly filling this space, with marketing research possibly becoming obsolete.

Implications for marketing education

How can we compete as marketing educationists? Obviously, we will not become more proficient than computer scientist in programming complex neural networks, artificial intelligence (AI), and machine learning algorithms in open-source software such as R and Python that are ran on Hadoop and cloud-based platforms. Although, the gathering of data and the consecutive analyses are of key importance this is nevertheless a small part of the Big Data and analytics revolution.

Most projects in commercial firms and governmental organizations fail, because the focal strategic issue is poorly defined or the results of the complex analytical techniques are not communicated clearly toward non-technical decision makers. Another bottleneck concerns the implementation of analytics results in business processes. Marketing research graduates could potentially contribute to overcoming these three main pitfalls, as the required skills concern a mix of technical and analytical knowledge, on one hand, and marketing and business knowledge, on the other hand.

In current business practice, much hope is directed toward visualizations when communicating the results of analytics to non-technical audiences. However, attractive pictures do not suffice; managers require at least a conceptual understanding of quantitative techniques and how analytics can enhance strategic and tactical marketing decisions. Moreover, analysts do not always have to be gurus in state-of-the-art machine learning or AI. For the analysis of transactional data, which are at the core of most brick-and-mortar companies, relatively simple analytical techniques that have been used for decades in marketing research perform as well as the later incumbents such as neural networks (Knott, Hayes, & Neslin, 2002). The basic concepts of the relatively simple techniques can be understood by most marketing graduates and by decision making. More relevant for the three main pitfalls, which were mentioned above, is that analysts should understand business decision making better than is currently the case.

The current state of analytics education in business schools

Given the business requirements discussed above, I propose that quantitative courses in the field of marketing should transform marketing students into either business savvy analysts or analytically knowledgeable marketing decision makers. However, Nunan and Di Domenico’s (2018) overview of UK universities shows that few suitable marketing research courses are available. The United States and the Asia-Pacific region are perhaps more developed herein. For example, North Carolina State University in the United States and The University of Melbourne in Australia have developed excellent programs that teach analytics from an applied perspective. They deliver graduates who can formulate practically applicable analytics-based advice for business decision makers instead of undecipherable computer-science models. Such programs are academically multi-disciplinary and connect strongly to practitioners through guest lectures, work placements, in-class assignments based on company data, and so on (Davenport, 2014).

Breaking down the academic discipline-based silos and demolishing the wall between business and academia are obvious requirements for an applied academic field such as marketing and business studies in general. Fortunately, contemporary businesses are open to such collaborations, as it provides them access to talented graduates with high-in-demand skills. Some academics are also increasingly seeking collaborations outside their own silos.

Thoughts on Big Data from an educational perspective

Above I mostly addressed transactional data and proposed that much of the required analyses on such data can be conducted by quantitative marketing researchers. When analyzing transactional data, it may also be useful to work together with computer-science graduates to experiment with new techniques and compare their performance with the conventional quantitative marketing techniques. However, most value is added by computer scientists when analyzing unstructured, high-velocity data such as web-surfing data, social-media data, data derived from video-recordings, and spoken voice-call data from the company’s customer helpdesk (Davenport, 2014).

These Big Data are often the only information source for online companies, but are also becoming increasingly important for brick-and-mortar firms. For the latter type of firm, Big Data has the potential to enrich transactional databases, after these unstructured sources have been transformed to a row-by-column format. The resulting data are added to the data warehouse that also consists of structured rows and columns (Davenport, 2014). Consecutively, relatively simple statistical techniques, such as cluster analysis and logistic regression, can be applied for gaining insights that are relevant for tactical and strategic marketing decisions. Such insights can, for example, support prospect selection for marketing campaigns, credit scoring models, retention models, customer segmentation, and so on. These applications can remain within the domain of marketing research, next to other traditional marketing topics, if we teach our student to work with transactional data and Big Data and enhance the connection between analytics results and strategic decision making.