As we searched for ways to deliver our first editorial, we decided to learn from our experiences as researchers. After all, we seek to discover new ideas and create new pathways from which to enhance the journal, so why not begin with a qualitative exploration? We chose to combine the purpose of the journal itself with the idea of stepping back to explore the term “analytics”; from this, we decided to utilize semantic analysis.

Exploratory study

The goal of the study is to gather qualitative information regarding the concepts and themes involved in and around marketing analytics. To do this, we chose to use an adapted version of netnography as our methodology, a technique which involves the study of online communities or material (Kozinets 2002).

Procedure and analysis

As a first step in understanding the issues, conversations, and ideas surrounding marketing analytics, we began with a multiple-phase netnographic analysis. In the first phase, in March of 2017, two trained graduate students who were unaware of the purpose of this research were tasked with gathering data about the term “marketing analytics.” Based on the literature, the students followed these instructions: (1) use the four terms: marketing analytics, business analytics, decision analytics, and process analytics; (2) For each term, perform a Google.com search; and (3) For each of the first few pages of search results, open each website link, identify the place where a definition for the term exists, and store it in a spreadsheet. This process led to N = 64 definitions, with n = 18 for “marketing analytics,” n = 17 for “business analytics,” n = 15 for “decision analytics,” and n = 14 for “process analytics.” We chose an unguided semantic analysis tool called Leximancer to explore themes from these definitions. Leximancer (www.leximancer.com) determines themes and their underlying concepts (subthemes) from qualitative data (Smith 2011) using a machine learning technique. Many existing research studies from various disciplines contain analysis performed by this tool, which is based on Bayesian theory (e.g., Campbell et al. 2011; Dann 2010; Kirkendall and Krishen 2015; Krishen et al. 2016). The tool also provides comparative diagrams derived from multiple files, such as comparing these four different search term definitions and calculating relative weights of themes and concepts from them (Rooney 2005).

Results

Analysis of the data consists of semantically characterizing each of the definitions of the search terms followed by a comparison of the aggregate results from each term and its interrelationship with the aggregate results from the other terms. Figure 1 provides a set of themes and their interrelationship, mapped with respect to the four search terms, shown as FILE_marketing analytics, FILE_business analytics, FILE_decision analytics, and FILE_process analytics. To define all four of the terms as a complete aggregate, as shown in Fig. 1 with large circles and provided in the left-hand column of Table 1, the outcome themes include data, analytics, analysis, predictive, product, time, customer, results, value, strategies, and impact.

Fig. 1
figure 1

Analytics themes and concepts

Full size image
Table 1 Analytics sample definitions (n = 64)
Full size table

Discussion

Sample definitions for each of the themes and the concepts within them are provided in Table 1. The themes are ordered from the most prominent (or most frequent) to the least frequent in the table. Alongside this Table, Fig. 1 shows the concepts within the themes and highlights the prominence of the concepts by various sizes of gray circles within each of the themes, which are represented by the larger circles. The eleven larger circles represent the themes, and the concepts are overlaid within them on top of smaller gray circles. These themes are connected to each other to represent the frequency of representation of the concepts with each other. The analysis also provides relational information about the concepts across the four different search terms (marketing, business, decision, and process analytics). The direct concept connections from marketing analytics include sales, customer, product, media, and strategies; from business analytics include analysis and performance; from decision analytics include models, and time; and from process analytics include organization. The terms connected directly to the analytics concept include organization, predictive, improve, identify, focus, and understanding. These ideas center on gaining an understanding and then identifying and focusing in an organized manner to provide a predictive solution.

Conclusion

Using our data-driven process to explore various definitions of analytics, we present Fig. 2 which proposes our conceptualization. In the center of the diagram are the four key terms that were searched and the outcome themes directly connected to them. For business analytics—analysis and predictive were the main themes, for process analytics—predictive, for decision analytics—data and time, and for marketing analytics—customer, product, and strategies. Thus, we view analytics as an interdisciplinary and inclusive area of study which aims to be insightful and ultimately enable timely and influential firm-driven campaigns. In this vein, we welcome contributions from multiple disciplines of scholarly discourse as well as timely practitioner insights and ideas to the Journal of Marketing Analytics.

Fig. 2
figure 2

Conceptual view of marketing analytics

Full size image