Discussion and outlook

The paper at hand contains a taxonomy and structured literature review to shed light on the BI & A 3.0 research developments of the past decade. Our research contribution is fourfold. First, we provide a systematically derived taxonomy for structuring Bi & A 3.0 research. Other researchers may apply this taxonomy to classify their BI & A 3.0 research results. Second, based on this taxonomy, we provide a structured overview about current IS research in the area of BI & A 3.0 between 2010 and 2018. The results clearly provide research topics of high interest as well as research gaps. The later may inspire other researchers to address them in further research projects. Third, we provide an answer to the MISQ special issue article of Chen et al., in which the authors predict upcoming big data research from the perspective of 2012. We partially confirm their prognoses, such as the trend of increasing publications of BI & A 3.0 over time. Fourth, we provide a research agenda that clearly points out open research issues that address dimensions and characteristics, which currently receive less attention in IS research. In total, we suggest nine research questions that are not answered yet.

Practitioners may use our results as a starting point to find a suitable data analysis solution for business challenges or even to find relevant expertise per taxonomy dimension. In particular, the eleven evaluated BI & A approaches could be valuable for practitioners as its implementation in organizations will probably be easier as non-evaluated approaches.

Furthermore, our results contribute to the controversial debate of Benbasat and Zmud, who identify areas belonging to IS and represent the core of the discipline. Since all papers analyzed in our literature review discuss its investigated phenomenon in the IS context, the results provide evidence that BI & A 3.0 is clearly part of the discipline. But BI & A 3.0 is also influenced by other disciplines, such as computer science, mathematics, economics. Thus, the boundaries are vague to assign the topic clearly to one research discipline. According to a Delphi study of Becker et al. leveraging "knowledge from data, with […] high data volumes" is one of the grand challenges of research in IS, which is one of the key characteristics of BI & A 3.0. Arguing with Becker et al. and Benbasat and Zmud, BI & A 3.0 can be perceived as a central part of the IS research discipline.

However, the results of this study are limited. We investigate the status quo of BI & A research in the IS discipline by analyzing the top IS literature. To define the scope of journals and conference proceedings, we apply the JOURQUAL 3 ranking for information systems and focus on A+, A, B, and TOP 30 C outlets. We did not regard other potentially relevant publications like those from other disciplines, lower ranked outlets, special-interest workshops, or industry reports. The paper selection might be biased, because we focused on papers considering IT/IS artifacts. Furthermore, we focus on the application of the developed taxonomy for structuring big data research. Thus, we did not regard other dimensions, possibly relevant for other scientific disciplines, in our literature review. Our results and findings are limited to the applied search string, which covers most, but not all, of the IS research in BI & A and could be possibly extended, e.g. with additional keywords and more synonyms. Some emerging aspects like the block chain technologies were not mentioned by Chen et al. and were not in scope of our investigation. The relation between research and practice remains ambiguous as the field is very dynamic and reports from practice like industry reports were out of the scope in our research. Possibly, the academic literature is lagging behind the current trends in practice as companies like Google are publishing a lot in the area of data science and big data analytics outside the traditional academic world.

Based on our research contribution, further research is needed to close the identified gaps in big data research. Our findings motivate IS researchers to address e.g. the fields of E-Government and Smart health. The limited scope of IS scholarly publications opens room for repeating the search for outlets in closely related disciplines, such as informatics, mathematics or economics. In addition to academic literature, BI & A related industry reports could be analyzed as they might offer useful insights about current and forthcoming developments from practice. For all further classifying works in the context of BI & A 3.0, we recommend applying and possibly extending the developed taxonomy, presented in this paper. The work at hand is a starting point for ongoing research in the field of BI & A 3.0 and may foster fruitful discussions on the further development of this emerging research area.