Business Intelligence (BI)

The Objective

The goal of the enterprise is to leverage the various sources of data residing within disparate business systems, as well as additional external sources of information, to improve customer experiences through tailored or personalized content, to improve products and services, and to improve efficiencies and lower costs. Some specific examples of predictive analyses would be deriving marketing intelligence to target ad campaigns or to identify customers at risk of churning; performing risk analysis on issuing contracts or loans or making buy/sell decisions; or detecting fraud or corruption. These functions sometimes need to be implemented as real-time continuous monitors. Businesses prefer to remain flexible as to where data resides, e.g., to take advantage of public-private cloud configurations, and also to be able to migrate readily to new business applications and new analysis methodologies as they become available. What they require is the ability to readily assemble “collective business intelligence” across their diverse and constantly evolving hardware and software infrastructures.

The Challenge

In today’s environment, it has become common for organizations to adopt a strategy of pooling data from various sources into centralized “data lakes”. Users from within the organization that need certain data must be able to access what they need, pull it from the lake, aggregate it, and mediate it into training datasets that can be used for learning their models.  It is becoming increasingly critical to provide governance of segregation and privacy. Data dynamics arising in the sources are continually dumped into to the lake, and enterprises with complex heterogeneous and evolving environments have ongoing upkeep and maintenance to contend with. While there is increasing interest in migrating enterprise data to the cloud, it is still commonplace to find hybrid configurations where certain data continues to reside on premises due to legacy systems, regulation, or proprietary considerations. The fact that the underlying infrastructure and applications and locations for data continue to evolve creates a constantly moving target. When the analytics (in this case: predictive decisions) are required to be delivered in real-time, entire families of data architectures may need to be designed as part of a comprehensive strategy, and validated in ongoing fashion as the business evolves.

The Collaborative Analytics Solution for Business Intelligence

Collaborative Analytics (CA) provide a way to bring together the collective intelligence locked up in distributed business systems, while avoiding data integration and making the mixed distributed infrastructure appear to the business analyst as a single system. It operates as a lightweight solution fabric to deliver predictive analysis while abstracting away the details of the underlying participating systems.  It is infrastructure-agnostic, rapid to deploy, and low cost-to-adopt. Because the pain of data integration is avoided, businesses are able to evaluate the use of more sources of data, and in different combinations, to deliver superior business intelligence at lower cost and latency.

Several benefits derive which are highly valuable in the context of BI applications:

· Reduced contraints on infrastructure and applications that provide businesses flexibility to move data and migrate applications at will, while maintaining an ability to do predictive analysis across that data; flexible and dynamic parsing of data between on-premises and cloud is enabled

· Ability to take immediate advantage of data residing in legacy systems, or data that resides in the (“incompatible”) systems of a new acquisition

· Accelerated analytics workflow which increases competitiveness by allowing businesses to iterate and experiment with key questions, support in the data, algorithmic techniques and different sources of data; this empowers users to hit tight windows of opportunity in the market

· Low cost of adoption to deploy and try a variety of analysis architectures; these may also be used in conjunction with data integration strategies to help optimize the overall strategy

· Scalable architecture that can handle many originating sources of distributed information (thousands+), making it unlikely that the spread of information will ever create a barrier to analysis or otherwise impede time-to-analysis-results

· Providing a path for businesses to supplement their internal data with external sources of information, including sources that could never be accessed or integrated due to privacy, proprietary or other restrictions

· Reduction in recurring costs for governance, redundancy, security, and maintenance of large central data repositories which become business critical

CA provides an agile and flexible distributed framework for “collective business intelligence” that is future-proof and can scale with the business.