Surely every major enterprise application will be re-designed to have “Big Data Inside”.
I recently wrote an article called the “Predictive Enterprise Awakens” and the more I’ve reflected on this, the more I’ve realised two core trends (that are bigger than any one company) that will drive big changes across business and the enterprise software landscape:
Big Data and Big Compute advances make it possible for many enterprise software applications to be re-imaged to embed “Big Data Inside” the applications.
Machine Learning coupled with a beautiful user experience will make the ‘predictive’ phase of Big Data more accessible to more businesses.
There is still a lot to be gained by traditional Business Intelligence, but it’s really the rear-view mirror of a business. Yes it’s helpful to know how many Xbox One's have been sold in Seattle, but knowing the answer to “How many Xbox One’s will we sell in Seattle?” will have a more profound and immediate impact on how a business behaves in a closer to real-time fashion.
An important implication of this is that the Machine Learning capability will not be a separate standalone activity in of itself, it will become a capability of every application. Big Data will be inside - think Intel Inside. Up until now most of our Big Data progress is still really focused on collecting stuff. The real benefit to business will be a scalable, reliable, repeatable ability to process data, observe and learn patterns, and predict what is currently unknown based on what the machine sees.
What is Machine Learning? Why should I care?
During dinner recently my son (19) asked me what is Machine Learning — we have exciting meals! I said to think of it as the equivalent of life experience in humans but for machines. It's the ability to take what we’ve learned in the past and apply it to perhaps an unknown or less familiar situation today. Machine Learning scales this behaviour massively and will become embedded in our enterprise applications.
Take the humble spreadsheet or core Financial Applications, today they just present data that we have to apply our intelligence to — although in my case this is a limited contribution! Imagine when the humble spreadsheet, or ERP application, or CRM application starts to learn from millions of variables and outcomes. With “Big Data Inside” applications will become inherently more intelligent as the machine implicitly learns patterns in the data, corresponding outcomes and derives evidence based insights.
Is this really feasible? What do we need to make it happen?
It’s pretty much almost here. In simple terms for the Predictive Enterprise to fully awaken we need our applications to be able to capture and explore the data, visualize it, model it, ask a question or questions, weight the answers, perhaps bring in other data, and repeat the process on a scalable basis. The trick is the number and frequency of iterations for the machine to learn.
Today we have some pretty decent solutions, such as the Oracle Big Data Appliance, which we (Qubix) are familiar with and see great potential. This type of power creates a closer to real-time data capture and analytics environment. The machine learning possibilities inside applications will drive new business models, new start-up opportunities, and a raft of business and society level benefits.