Analytics tools, just like many other digital technologies, shift from monolithic industrial mode to mosaic digital mode.
Predictive Analytics (PA) is about applying quality data to forecast what will happen in the business. Analytics democratization as "making PA easy to use for many new user groups" is a good thing in many respects: First, it is just mandatory for Big Data becoming true mainstream. It also will help to increase the amount of domain level expertise that will be brought to use Big Data, and therefore speed up discovery of new facts in many domains, and also this democratization will help a new class of solutions, tools, and services to find a market. More specifically, what are the characteristics of the democratization of Predictive Analytics?
Intuitive UI: With a good intuitive UI and clever use of algorithms in the back, you can enable a very good working environment for domain experts with some general reporting and statistics skills. In that sense, you will see advanced analytics capabilities becoming additional features.These tools will make it easier for business users to do "basic" analytics more easily and with larger and more complex data sets, etc. But this will only serve to grow the market for predictive analytics in general, which will mean more work for data scientists doing more heavy lifting in different areas. If you define a stable problem space for such an environment, could you even solve some of the feature generation and data preparation tasks for those users as part of an initial setup. This will limit them to work in a predefined problem space, which of course reduces the exposed complexity.
Analytic accretion: Some level of "analytic accretion" occurs in various verticals or domains to facilitate a sustainable open framework for predictive analytics; each will have to tune or determine prospective value. Predictive Analytics has interesting temporal parameters in achieving outcomes in verticals. It's going to get real interesting to see where the heavy lifting is going to occur. You will see more business user facing Advanced Analytics functionality, and that can be a good thing as long as you have a very defined problem space, prepare the respective data beforehand and focus on user driven result interpretation and usage instead of process automation.
Self-service: Self-service BI and reporting is easy as well, if you have a good data model in place already. There is no Predictive Analyses without good data. It has never been so complex to achieve data integration and data quality as it is today where companies need to mix relational and semi/unstructured data. But talent is still the tool master. Instead of statisticians running out of work, it only becomes more clear that it was not the tools made the analyst, but the understanding of what the tools were doing. And better analysis tools allowed talent people to put their understanding of tools for better and more efficient use.
Analytics tools, just like many other digital technologies, shift from monolithic industrial mode to mosaic digital mode: from heavy to light weight; from full set of enterprise software to nimble consumerized application, from hard to use to user friendly, and the democratization of predictive analytics well reflects the nature of digital transformation, to make things better, faster and cheaper; and to make the work more purposeful, innovative and smarter.