Search results

(1) Students writing their theses need to address specific research questions, and take clear positions with respect to these questions. (2) They should strive to present, validate and discuss their own research contributions, and in so doing demonstrate how they have made distinct contributions to knowledge. (3) An appropriate structure for this purpose is to start at the end, by writing the thesis around key findings. (4) Throughout, the aim should be to communicate effectively, so that the evidence and position with respect to the research questions is crystal clear and unambiguous. (5) Academic or institutional conventions which impede these goals should be firmly challenged.

AbstractDashboards are commonly used to inform data‐driven decision‐making by multiple stakeholders within and across businesses. The purpose of this article is to show how a comprehensive marketing model, the NBD‐Dirichlet, can be used to construct coherent and integrated dashboards. This is demonstrated using an example that offers practical guidance to practitioners and researchers for incorporating the model into a dashboard and showing how it can enhance visualisation, communication, and decision‐making. The example concerns digital media consumption behaviour, specifically the section choice behaviour of readers of an online magazine. The example demonstrates the utility of the NBD‐Dirichlet model to underpin a marketing management dashboard (RQ1), where model parameters are estimated from unstructured log‐file data (RQ2) using log‐likelihood estimation (RQ4). The example also shows the applicability of the model in analysing a non‐brand attribute, specifically magazine content sections (RQ3). From inspection of graphical and tabular dashboards, it is evident that magazine section content is read by consumers in ways we might expect given the well‐known Double Jeopardy (DJ) pattern of the NBD‐Dirichlet model (RQ5). There is no evidence of change‐of‐pace behavioural loyalty (RQ6), nor niche behavioural loyalty (RQ7). Finally, the article highlights the benefits of the NBD‐Dirichlet in business as not only a tool for underpinning dashboards but also for scenario planning (RQ8).

A central theme in brand choice is the modeling of the drivers of repatronage behavior. The work reported here focuses on a particular service context—airline repatronage. The distinct problems of model specification in this context are discussed. Various models of airline repatronage are estimated using random effects probit modeling. It is found that although performance perceptions and punctuality of the airline have a role to play, in particular in terms of choice on the next purchase occasion, past purchase behavior is the key driver of repatronage. This is evident across the various models that we examine, suggesting that the findings are robust. Implications of the modeling approach and the findings are briefly discussed in relation to the airline context and service marketing more generally.

An empirical generalisation is a relationship between two or more variables that has been observed across a range of conditions. By knowing that an observed relationship holds under a range of conditions (and that it does not hold under other conditions) it is possible to use knowledge of the relationship for practical purposes, such as making routine predictions and stating principles. It is also possible to start to theorise why the relationship occurs, and why it holds under some conditions and not others, thereby moving from empirical description to theory-building. The importance of this form of knowledge for marketing is examined. Practical measures are suggested to encourage the search for empirical generalisations.