Workshops held at the University of Hull in July 2014 and the OR56 Conference in August 2014 were used to identify the challenges organizations face in creating value from big data analytics. From the workshops thirty-one challenges were identified.
A Delphi study was then conducted to order the thirty-one items in terms of their perceived significance as challenges in creating value from big data analytics. Round one of the Delphi study was launched in October 2014 (72 respondents) and round two in November 2014 (42 respondents). Sufficient convergence was achieved in round 2 to close the study. The respondents were a good balance of practitioners (those working for organizations), consultants, and academics.
Respondents were asked to rank their top 10 items (rather than attempt to order all 31). The top 10 items are:
- Managing data quality [assuring data quality aspects, such as accuracy, data definitions, consistency, segmentation, timeliness, etc.]
- Using analytics for improved decision making [linking the analytics produced from big data with key decision making in the business]
- Creating a big data and analytics strategy [having a clear big data and analytics strategy that fits with the organisation’s business strategy]
- Availability of data [the availability of appropriate data to support analytics (does the data exist?)]
- Building data skills in the organisation [the training and education required to upskill employees in general to utilise big data and analytics]
- Restrictions of existing IT platforms [existing IT platforms/architecture may make it difficult to migrate to and manage big data and analytics]
- Measuring customer value impact [can the real impact on the customer of managing big data be measured?]
- Analytics skills shortage [difficulty in acquiring the mathematical, statistical, visualisation skills for producing analytics]
- Establishing a business case [can ‘tangible’ benefits of big data be demonstrated (e.g., return on investment)?]
- Getting access to data sources [accessing appropriate data sources to produce and manage big data (can the data be accessed?)]
Further analysis is being conducted and the results and implications of the study are being written up in a working paper.