LESSON - When Business Intelligence Becomes Operational
By Gerald D. Cohen, CEO, Information Builders
When the concept of business intelligence emerged, it was often viewed as icing on the cake. While transaction systems were considered necessary for running day-to-day operations, BI systems were initially intended to provide historical insights on how or why the company was performing as it was. “Back-office” BI, performed by an elite group of business analysts at division or corporate headquarters, has certainly helped thousands of organizations slice and dice corporate performance, enabling them to identify costs, profits, or opportunities for improvement. Nonetheless, BI has traditionally been a reactive solution that was the traditional domain of a company’s “rocket scientists.”
Organizations arebenefiting from a newform of ‘operationalanalytics’ that areproviding real-timerecommendations andperformance indicators.
—Gerald D. Cohen, CEO, Information Builders
Recently, we’ve seen a growing cross section of organizations proactively working BI into everyday operations. These organizations are benefiting from a new form of “operational analytics” that are providing real-time recommendations and performance indicators, so action can be taken early to correct problems.
For instance, at Moneris, Canada’s largest credit card processor, merchants who subscribe to the company’s services are now receiving daily analytic reports benchmarking their performance against regional or product segment averages. Meanwhile, Ford is cross-purposing data from its engineering system to provide instant trend analyses on warranty and repairs to over 30,000 users across its network of affiliated dealerships. And a leading semiconductor manufacturer is applying the results of quality inspection analyses of wafer samples and converting the information into visual, geocoded maps using ESRI mapping software. That’s helping production line operators immediately visualize the results of their work.
With BI evolving into an operational system, several lessons became apparent. The first is that the boundary between historical and current transaction systems was dissolving. The result was not a debate between historical or real-time BI. Instead, the solution had to include delivering the right blend to address the business problem. For instance, if you are staffing a call center, you will need near-real-time information to gauge traffic patterns within the last hour, plus historical data to understand whether current activity is an anomaly. At the other end of the spectrum, verifying the accuracy of your company’s forecasting process requires historical data to paint the big picture.
Another piece of the puzzle is integration. While traditional, back-office BI applications were often narrowly designed to address a single or small defined set of business issues, operational BI systems by nature tend to be more broadly based. This reflects that a single source cannot provide the big picture on issues such as how well a company is serving various sets of customers. As BI shifts toward an operational focus, it draws on multiple sources and reports. Consequently, integration has to be part of the solution, not an afterthought.
In addition, when BI goes operational, it tends to attract a broader user base compared to traditional back-office data marts. Therefore, ease of use becomes a major goal when developing operational BI systems. As you address larger populations, you have to accommodate people who are not computer experts. Operational use is also likely to increase the number of reports that are developed, maintained, and modified to reflect changing business conditions. Consequently, reporting must become simpler for administrators, developers, and analysts to create and maintain.
In 30 years of serving customers, we’ve learned that there is no such thing as a typical reporting environment. Data sources, competitive requirements, audience sizes, and requirements for timely data for your company will differ from somebody else’s. Your BI solution must be flexible, powerful, simple, and scalable enough to address each of these needs. And it must be ready to step up to the plate when you need the operational analytics that can help you make the right decisions in the course of daily operation.
This article originally appeared in the issue of .