According to a recent survey of over 800 business and IT professionals, 91% of respondents said it is important for business users to have access to their data without the need for IT involvement. However, this is contrasted by the statistic that only 22% of those business users actually have access to use self-service BI tools when they need to do so.
Having discussed this staggering demand gap with a colleague, two trends were identified that contribute to the existing adoption rates of Self-Service BI.
Firstly, IT has overall responsibility for security and compliance. This creates a reluctance to devolve ownership of applications and data. There is a fear that the democratization of data brings 'anarchy' and the 'Attack of the Data Silos 2' (Noyes, 2015). Although this may induce a new wave of analytical capabilities for users, it also encompasses the issues of data-quality and duplications, lack of compliance, and the rise of data silos, wherein information and insights are not shared with other parts of the organisation. Furthermore, with only 10% of self-service initiatives being governed adequately, IT processes are being bypassed, causing IT departments to want to retain autonomy and control.
Secondly, there is a technology and user-skills gap that is preventing rapid adoption of self-service BI. Vendors need to focus on building tools that fit within the context of business users day-to-day work, and augment, rather than detract from their role within the organisation. Equally, business users share some of the responsibility and will be required to engage with training to bridge the skills gap and for companies to see a full return on investment. Therefore, until an equilibrium can be found between the tools available and the proficiency of users accessing them, we will not see the full value of self-service BI exploited.
In theory, self-service business intelligence (BI) applications are supposed to make organizations more agile by getting IT out of the way of business users. In reality, most organizations thus far have shown a limited ability to execute on those BI ambitions.