With complex event processing tools, data virtualization software, business process integration products and other advanced and still-emerging technologies at the heart of real-time data analysis systems, deployments require a certain level of sophistication to tackle the technical challenges that lie ahead.
And beyond the technology-related hurdles there are significant political and process issues that can trip up companies on real-time analytics projects, impeding their ability to deliver operational intelligence that helps improve decision making and business performance. Avoiding those pitfalls requires a well-choreographed implementation plan that gets IT, business intelligence (BI) and business managers on the same real-time page.
At the outset, many organizations get off on the wrong foot by pursuing real-time BI and analytics simply because it's a widely hyped technology, without first going through the process of building a formal business case. "A lot of times you have IT in charge of what's going on, and they make the call to go real time or near real time because they can," said John Myers, an analyst at research and consulting company Enterprise Management Associates Inc. in Boulder, Colo. "They're not asking, 'Does this technology fit the business case?' They're saying, 'This is really cool and we should do it.' "
Blind pursuit of technology for technology's sake can also mask some of the inherent complexities associated with real-time analytics. One of the most important issues to figure out, Myers and other analysts say, is ensuring that an operational system can support real-time analysis capabilities without any performance degradation on the transaction processing side.
More in this series on real-time business analytics
Learn how real-time BI systems are feeling the strain of dealing with big data
Read how real-time access to BI data informs decision making in business operations
Missing out on real-time BI requirements
Companies often also underestimate the data integration requirements of a real-time system or don't fully understand the scope of what's needed from both a technology and a project planning and management perspective, according to Lyndsay Wise, president and founder of WiseAnalytics, a midmarket BI and data visualization consultancy in Toronto.
"A lot of times smaller organizations don't end up budgeting for the different components," Wise said. "They want an analytics platform and set aside a certain amount of money, but they don't realize there's a data management side, a data delivery side as well as a data integration side." As a result, she added, "they end up not setting proper expectations" on the overall project cost.
As analytics moves directly into the operational realm through real-time initiatives, business expertise and involvement become even more important than they are on traditional BI projects.
"You need to have operational training and knowledge of how the business runs to do [real-time analytics]," said Colin White, president of consulting company BI Research in Ashland, Ore. "Just as the dividing line between what's operational and what's analytics is becoming gray, the dividing line for people in IT has to be gray." The key to success often is removing the line between business operations and the BI team altogether, White said: "Otherwise you run into conflicts about technology selection and budgets."
Too much information gets in the way
Collaboration between operational managers and workers and the BI team is also crucial to building business intelligence dashboards that make sense for the job at hand, said Roy Schulte, an analyst at Gartner Inc. in Stamford, Conn. BI developers "sometimes design real-time business dashboards as if they were designing a daily or weekly dashboard or report," he said, resulting in cluttered screens with more detailed information than operational users need to make minute-by-minute decisions.
Instead, dashboards for real-time data analysis applications should be designed "to convey a small set of urgent information quickly" while giving users the opportunity to drill down into the underlying transaction data when necessary, Schulte said. He also recommends that the dashboards be complemented by an alerting mechanism that sends email messages to the appropriate users when a real-time system detects business problems or opportunities.
Another big factor in making real-time analytics deployments work is ensuring that people and processes are in place to deal with the information and analytical findings being generated so the initiative can have a positive impact on business performance. Typically, that involves rethinking business processes, retraining workers and changing operational procedures to make an organization more responsive to customer purchasing decisions, fluctuations in financial markets and other business developments.
In fact, if business units aren't in a position to act on real-time intelligence, or if operational systems aren't fully equipped to handle real-time data feeds, a company will find itself investing money and resources in a program that has little chance of getting off the ground, warned William McKnight, president of McKnight Consulting Group in Plano, Texas.
"You don't want to let the cart get ahead of the horse," McKnight said. "If you haven't developed analytics [processes] to the point where you can really do something with the information and make an impact on the business, the use case is diminished."
Beth Stackpole is a freelance writer who has been covering the intersection of technology and business for more than 25 years for a variety of publications and websites.
Follow SearchBusinessAnalytics.com on Twitter: @BizAnalytics_TT