According to a recent Forrester survey, industry leaders invest 38 percent more of their IT budgets in business intelligence (BI) than their competitors, and nearly half of these companies obtain a double-digit return on their BI investments. Clearly, large investments in business intelligence pay off, but what do these BI leaders do that the rest of us can learn from? That was one of the questions that we sought to answer during a recent Innovation Series webcast, in which Forrester’s Boris Evelson joined me to discuss an intriguing topic: “Agile BI Without Shadow IT.”
In Boris’ view, success with BI has a lot to do with closing the gap between business and IT. IT professionals sometimes overemphasize the importance of having a centralized, highly optimized data architecture. While that is a worthy, long-term goal, it may not be the immediate objective. Most business people are looking for quick results, so establishing a single version of the truth is not always their top priority. They’re often willing to start with “good enough” and work on mastering their data later.
While technology professionals used to be responsible for developing BI content, and business users simply consumed that content, this situation has changed in recent years. Today, business users in marketing, HR, finance, and many other lines of business often develop their own analytics. As shown in the diagram (Fig. 1 "Three Patterns of BI Usage"), technology professionals may only be responsible for the tip of the BI pyramid, where the enterprise BI applications reside. More and more, the bulk of BI assets are being developed by business users, and shared with their peers:
Technology professionals are still responsible for developing complex BI content and overseeing BI governance. The business community needs their help identifying data sources, and this is an important role that IT professionals continue to play. (Some of these workers are reorienting themselves as data intelligence professionals.)
As shown in the right-hand side of Fig. 1, there are three primary types of data sources that today’s user community depends on: curated, modeled, and un-modeled. Curated and Modeled data includes enterprise data marts and data warehouses that have been cleansed and integrated to include prebuilt views and calculated metrics. Un-modeled data, which is often stored in flat, un-normalized tables, is useful for analysts and data scientists that need raw, un-curated data for data mining, data exploration, and data discovery.
Forrester’s “divide and conquer” approach depicted in the diagram meshes nicely with Information Builders’ core strategy of delivering self-service analytics for everyone, illustrated above in Figure 2. InfoAssist+ supports business people who need to source their own data, visualize that data, and manipulate it in unique ways. In addition, many power users depend on InfoAssist+ to create interactive InfoApps and share them with the business community.
Meanwhile, IT developers use WebFOCUS and AppStudio to create sophisticated BI and analytics applications:
This division of labor serves operational workers who want interactive, self-contained BI content that can be embedded within operational systems, as well as business analysts and data scientists who want more autonomy and control. It’s a great way to instigate an enterprise BI strategy and promote analytics to lots of people.
Finally, Boris pointed out that today’s business community needs to be able to access data from multiple sources, from ERP systems to social networks. For example, a marketing manager might want to combine social media data with sales data to gauge customer sentiment or determine how well a marketing campaign is working. Using a self-service InfoApp, she can peruse and visualize these data sources without having to wait for the IT department to develop something for her.