Understanding Users’ Action On Data Analytics Recommendations

Abstract

Current trends indicate that many organizations are making significant investments in Data Analytics (DA) to leverage big data. However, recent studies also indicate that a large percentage of these investments are unsuccessful and that a majority of users do not act upon a data analytics tool’s recommendations. This research draws upon the S-O-R framework and the Theory of Planned Behaviour to develop and empirically validate a theoretical model of the factors that influence/hinder a user’s concordance with and actions taken with respect to a DA tool’s recommendations. The model reflects the nuances of DA tool use within organizations including: (i) technological characteristics, (ii) individual characteristics, (iii) situational characteristics, and (iv) task-related characteristics. In addition, this study investigates the factors that shape a user’s perception of the quality of a DA tool’s recommendations, while trying to understand how, and to what extent, this perception influences a user’s concordance with, and the actions taken in regards to, a DA tool’s recommendations. The results of this research confirm that personal concordance and recommendation actionability are positively associated with user action on a DA tool’s recommendations. Moreover, perceived risk of action was found to be negatively associated with user actions taken with respect to a DA tool’s recommendations. It was also found that DA tool recommendation quality is shaped from intrinsic data quality, contextual data quality, DA tool quality, DA tool recommendation understandability, and analyst competency.