Big data analytics solutions: The implementation challenges in the financial services industry.

摘要

The challenges of Big Data (BD) and Big Data Analytics (BDA) have attracted disproportionately less attention than the overwhelmingly espoused benefits and game-changing promises. While many studies have examined BD challenges across multiple industry verticals, very few have focused on the challenges of implementing BDA solutions. Fewer of these studies have focused directly on the financial services industry, and none have quantifiably measured the severities of the challenges. That created gaps as BDA solution implementers in the financial services industry could neither access a roadmap to guide their steps against obstacles that lay ahead, nor compute the severities of the challenges. This study addressed those gaps through two research questions: (1) What are the challenges of implementing BDA solutions in the financial services industry; and (2) What are the rankings of these challenges, in terms of importance and relative severities, such that BDA implementations can devote more research attention to or hedge better against those challenges? To answer these questions, the study used a mixed methods approach to content-analyze 75 BDA documents and collate a comprehensive list of 22 BDA challenges. As well, 36 financial services industry BDA-subject-matter-experts (SMEs) were surveyed to validate the list, rank the challenges, and measure their impacts. The research findings showed that the challenges of implementing BDA solutions in the financial services industry are mostly strategic and people-driven, rather than process-induced or technology-driven. Specifically, miscommunications and misconception of the meanings, intents, and the value-added benefits of BDA implementation in the financial services industry were found to be the top challenge. Details of the results, its implications for the BDA communities of practice and discourse, and opportunities for future research were discussed. The results can be generalized if scaled with a bigger sample size and better measures-of-intangibles.