Gartner’s 2015 Magic Quadrant for BI and Analytics Platforms reveals a rising interest in tools that provide self-service data discovery capabilities. These tools give non-technical or semi-technical business users the ability to explore data coming from different sources and detect trends and patterns within this data, with the ultimate goal being improved decision-making and reaching practical insights for more efficient business processes.
The prominence of data discovery platforms is part of a broader trend in business intelligence: from retrospective reporting meant to evaluate past actions, to predictive and prospective analysis aimed to affect the way organizations reach decisions beforeDescriptive → Predictive → Prospective Analytics
To elaborate on the differences between the three one needs to understand the scope in which each type of analysis is performed and what it aims to achieve.
In the not-so-distant past, business analytics were mainly used a decision support or decision evaluation tool. For example, the business decided to invest resources in a certain market; at the end of the quarter, the data is gathered and the soundness of this investment is evaluated based on sales, ROI, etc. This type of analysis is always focused on past actions, although of course it aims to provide insight which could be relevant for the organization’s future moves.
This is a more advanced form of analysis, and would usually require more complex knowledge of data science. Predictive analysis attempt to make an ‘educated guess’ as to the probabilities of certain developments, based on past data. Essentially it is an attempt to predict the future’s data based on currently available information.
The next stage of predictive analysis, prospective analytics goes one step further: it not only aims to predict developments and emerging patterns, but to actually offer possible courses of action and estimate their probable outcomes. Naturally, if this vision can be translated into a reality, it can deliver immense value for any organization at a decision-making crossroads — by eliminating guesswork and reliance on hunches and gut instincts, replacing them with cold hard numbers and probabilities.
What Does this mean for the Industry?
The archetypal product of descriptive analytics is the static report, which summarizes the examined subject matter for business executives. This was the the common result of most business intelligence endeavors in the past. Since the analysis was retrospective in nature, it was acceptable that these reports could take days or week to generate.
But with the increased emphasis on forward-looking – i.e., predictive and prospective analytics, static reports are becoming less significant, with the focus increasingly shifting to interactive dashboards. These are better suited for the cause as they enable real-time data exploration and discovery, which are the building blocks of any predictive or propsective analytics project.
This also means organizations will be looking for tools that can give them answers faster, which at the same time will include the advanced statistical and mathematical functions required for creating complex predictive models. In many cases this means resolving some of the traditional bottlenecks of business intelligence, including the technical threshold that traditionally meant only IT or dedicated BI teams could use the tools at hand. Hence, it is likely that we will continue to see more products geared towards self-service BI and automated data preparation, in addition to features such as R integration to let data scientists really ‘go crazy’ with the data they have.
Alas, this is merely an educated guess. Surely a well-done predictive analysis would have produced a better one…