How Proactive Analytics Can Improve Your Business
In 1999, sociologist Neil Gross predicted, “In the next century, planet earth will don an electronic skin. It will use the Internet as a scaffold to support and transmit its sensations.” Today, in the second decade of this ‘next’ century, IoT, Cloud Technology and Big Data Analytics are knitting the very fabric of that ‘electronic skin’ while it is simultaneously ‘textured’ by technologies for artificial intelligence and virtualizations.
The rate of data generation in the IoT ecosystem would rise exponentially as the technology further penetrates in the market and masses. Gartner predicts that the end of the decade, the ecosystem would comprise of more than 26 billion units excluding PCs, tablets and smartphones. The research firm also projects that by 2018, fifty percent of IoT solutions will be provided by startups which are less than three years old.
Corporate benefits of IoT does not have to be confined to mere tracking, measurement or generating insights from previously aggregated data warehouses. It is seen that organizations are not inclined to tapping insights from proactive analytics: to define a course of action (prescriptive analytics) or predict outcomes (predictive analysis) based on streaming analytics (from data generated near real time). A study by Forrester points out that companies analyze only about 12 percent of the data harnessed by their IoT tools. On the flip side, personal service providers rely much on proactive analytics. The ability of Spotify app to choose music in tune to the pace of the listener is a good case in point.
Proactive tools may comprise of applications/APIs installed on devices which is capable of sending data in a pre-processed form (predictive data). This in particular would aid in simplifying predictive analytics and it also turns out that the volume of such predictive data is significantly lesser and thus enables quicker processing especially for time-series data. Yet several experts are not entirely convinced with the ‘on device’ approach owing to the processing power it may demand. The IoT friendly Arduino platform for instance, is currently not equipped with adequate processing capability in their default configurations and installing proactive wares could affect its performance. This is a visible hurdle as IoT embraces miniaturization, yet we could bank our hopes on the underlying principle of Moore’s law.
In several cases of corporate BI, data appears to be perishable or in other words time bound in terms of value. Depending on the process and vertical, the time windows may range from microseconds in the case of financial trading to milliseconds and half a second of latency for fraud detection and recommendation engine respectively. Firms would be able to eliminate probable customer service delays while being able to address shortcomings with an extra edge. Consider the example of a GPS based coupon generation system, what good would it do when the customer drove past the coordinates of space? “Perishable insights can have exponentially more value than after-the-fact traditional historical analytics,” says Mike Gualtieri, principal analyst at Forrester Research Inc. “Most companies are not capturing perishable insights”
Considering the skin and the way the sense of touch works, signals are either interpreted by the brain or the spinal cord (as in the case of reflex action, for a ‘quicker’ response). An interesting analogy here in the context of IoT is—if in the cloud computing, where the stored data warehouse serves as the brain, Proactive tools on the cloud indented for processing streaming data would be similar to the spine.