Big Data processing improvements over the last several years have been incredible. Between in-memory solutions and massively parallel high performance analytics, it is just mind-boggling the computations that can be completed almost instantaneously. With the near universal connectivity and information available from both Big Data and transactional data sources, it is now possible to get the answer to just about any question no matter if it is from transaction histories, social media interactions or machine to machine sensors. It all translates into faster problem resolution and greater insight into customer satisfaction, which in turn should help every business create higher value for their customers.
However, enterprise applications have not maintained pace. Not only are they poorly suited for mobile, a topic in one of my previous posts, but today’s enterprise applications are woefully inadequate at integrating all that data into business users’ everyday workflow so they can react ahead of time.
The solution is not more middleware or enterprise application integration, as some of the big stack vendors such as IBM, Oracle and SAP have suggested. Instead, the solution is to divide and conquer, so to speak.
First, keep IT focused strictly on the data, keeping it clean, keeping it secure, making it available from every source (including Big Data sources), providing metadata views, etc. Second, permit business users to create their own processes by using their own wizards that leverage this newfound intelligence over the data.
In this new enterprise architecture, data services, created by IT, become the foundation for how we do our work at our desk or on the road. Each product group, geography and functional area would then create their own “customer centric” methods of interacting with the data and not be held hostage to antiquated applications that need to be rewritten or optimized.