Pervasive Business Intelligence
Historically, business intelligence and data warehouses have been associated with back office employees. In the 1990s, data was integrated and loaded nightly, optimized for trend reporting and strategic analysis, and delivered via reporting tools. Back office planners running pre-defined reports were the bedrock users of business intelligence (BI). Over time, knowledge workers evolved to demand richer, more diverse insights. As usage matured, requirements to include predictive analytics, event-driven alerts, and operational decision support have become the norm.While demand for near real-time information always existed in front office operational communities, the costs and complexity of loading data multiple times per day kept data out of reach. This caused expert database administrators and innovative software vendors to find ways of reducing data latency between source systems and the data warehouse (DW) to serve up “just-in-time” insights throughout the business day.
For many, the dividing line that kept (DW) information isolated from the front office vanished. Going beyond reporting, visionary enterprises found ways to enhance revenue and cost efficiencies across multiple business process and knowledge user communities. Front-line employees, suppliers, consumers, and business partners are finally being offered BI capabilities for tactical decision support. In this way, pervasive BI emerged within enterprises that compete using data analysis.
Pervasive BI Applications
Just as important as the executives who formulate corporate objectives are the rank and file employees who execute those objectives. Thousands upon thousands of small decisions made daily by employees, customers, and business partners add up to profitability and brand image1. Yet many front-line workers have historically made decisions without access to the necessary facts to achieve alignment with corporate objectives. The goal of pervasive BI applications is to take the data that produced the back office ROI of more than 400%2 and deliver it to front-line employees in a form appropriate to their job functions – with similar results. There are thousands of business process steps in a typical enterprise where pervasive BI insights can be added.What are leading companies doing with pervasive BI? Here are a few highlights:
> Inbound and outbound cross selling by Customer Service Representatives
> Point-of-sale fraud detection
> Automated insurance claims triage and fraud detection
> Personalized next best offer on web sites, ATMs, and POS devices
> Supply chain business activity monitoring
> Retail out-of-stock monitoring on promotional items
> Labor and crew scheduling
> Optimizing trailer and container loading and routing
> Anti-money laundering
> Dynamic pricing and yield management
> Real-time manufacturing line quality alerts
Pervasive BI injects new workloads into the major BI subsystems. The classic extract-transform-load (ETL) subsystem now must cope with near real-time integration of data from multiple continuous data sources. The DW must support a mixed workload of constant data loading, complex reporting, and tactical requests. Subsequently, the BI platform must analyze all the data and deliver timely information through reports, alerts, dashboards, and operational applications. A new information supply chain must be established to deliver pervasive BI solutions.
Pivotal Role of the CIO
The hardest part of deploying pervasive BI is changing the way an enterprise thinks about BI. The CIO must be a change agent – a thought leader in evangelizing the possibilities of a new vision and the required capabilities. By reaching out and educating enterprise executives and line of business owners, the CIO helps the organization conceive new ways to increase revenue or reduce costs using low latency information. This usually begins with the question, “What would we do differently if all employees had integrated analytic information in minutes instead of overnight?” The first step often involves educating the organization about how other enterprises have gained competitively from pervasive BI solutions.
Second, the CIO must convince the IT architects and developers that pervasive BI can be implemented successfully. Many technical skeptics will surface, often with a vested interest in older architectures and designs.While the skills needed may exist within the IT organization, the J2EE/.NET developers and BI architect communities have rarely collaborated in the past. The show me skepticism of legacy developers must be satisfied to move forward. CIOs are accustomed to the challenge of organizational change, but that doesn’t make it easy.
