Decentralizing data while making it “less dumb”

What to sell and to whom? How to build product most efficiently? When and in what quantities? Each of these tasks is difficult; reaching the optimal outcome across all of them is almost impossible without relevant, timely information. The path to such information, however, is complex, time-consuming, expensive, and not always successful. Before the organization can even start analyzing its data, it generally must be assembled by importing, scrubbing and normalizing from multiple disparate sources, in different formats, using different data models, etc. Current data-management systems and approaches assume that data is “dumb” and that data must be managed using sophisticated and complicated infrastructures. Organizations spend millions of dollars building and implementing fragile data hierarchies with multiple point-to-point connections. Data is stored in centralized data warehouses and lakes spanning large server farms.