Most organizations have traditionally focused on using business intelligence systems to enhance strategic and tactical...
decision making, but in recent years, a growing role has developed for BI tools that can provide real-time analytics capabilities in operational processes. Now a question for BI managers and business executives to ask themselves is: Has operational BI become a necessity in their organizations?
There is a certain amount of urgency around big data; we know that the volumes are exploding, and yet many organizations are missing the opportunities of big data.
research director for data management, TDWI
In a webinar organized by The Data Warehousing Institute (TDWI) last week, Claudia Imhoff, president and founder of Morgantown, W. Va.-based data management consultancy Intelligent Solutions Inc., said she thinks operational business intelligence (BI) is something all companies should engage in to stay competitive in today's business climate. Embedding BI functionality into business operations can smooth out the supply chain, improve employee productivity and speed up the decision-making process at all levels of an organization, she said.
The idea is to take real-time or near-real-time data, analyze it as quickly as possible, and get information in front of operational workers right away so they can make more-informed decisions. For example, Imhoff said a manufacturing facility might monitor streaming sensor data to ensure machinery is operating at optimal levels; marketers could closely follow social media posts to identify situations where customer service representatives might need to respond to dissatisfied customers; and shipping operations could track the location of vehicles to make sure packages will arrive on time.
But Imhoff warned that businesses need to know what they're doing before they dive into operational BI. Conventional data warehouse and BI systems might not meet the real-time data collection and processing demands of operational decision making. As a result, she said, companies may need to start by implementing something other than a standard relational database to handle the large volumes of data that can be generated by machine sensors, social networks, financial information services and other sources of streaming data.
"Many times these are the NoSQL, non-relational databases [or] Hadoop -- whatever it is we can use to get screaming performance on large amounts of data," Imhoff said. She added that getting the right database in place is one of the most important technical considerations for BI teams because it determines whether data analysis can truly be done in real or near real time.
Businesses also need to think about how bringing BI technology into operational applications will affect employees, particularly front-line staffers such as call center workers and shipping managers. In many cases, those workers will have to change their decision-making processes to effectively utilize the information they get from operational BI systems. In addition, they are likely to be less tech-savvy than business analysts and other users of regular BI applications, Imhoff said. Therefore, managing implementations is not just a technology issue; she recommended a sharp focus on educating end users about the new system before it goes live so they fully understand how to use the data it generates and how the new capabilities change their job expectations.
Read more about adopting operational
Learn how real-time analytics can speed up business decision making throughout an organization
See why enabling operational and real-time BI processes takes both IT skills and business savvy
Get more advice on how to manage an effective operational BI deployment
In a separate TDWI webinar last week, Philip Russom, the organization's data management research director, said organizations that aren't leveraging the growing volume of streaming data to improve operational decision making are missing a huge opportunity, especially with big data applications. "There is a certain amount of urgency around big data; we know that the volumes are exploding, and yet many organizations are missing the opportunities of big data," Russom said.
One reason for the reluctance to embrace operational BI is that traditional BI and data management tools often can't handle the large volumes of data and the real-time analysis it demands, according to Russom. When evaluating different products, he said the most important thing is to make sure they really can analyze up-to-date data and deliver reports in real time, which he defined as seconds to milliseconds.
"That's a pretty steep technology requirement," Russom said. But, he added, it's the kind of performance "we really should expect from good analytics" in operational processes.
When looking for potential operational BI uses, Russom recommended finding business processes that have some sort of time-sensitive element. Examples he cited included online retailers monitoring customer browsing patterns that are indicative of potential churn, and manufacturers monitoring sensor data to spot equipment that might be on the verge of failure. In each case, he said, the end goal is using timely information to head off problems before they occur.