Written by Kathleen Lau, ComputerWorld Canada
While business intelligence (BI) has migrated to the position of corporate asset from back office application, a typical BI environment remains nonetheless complex and, according to a Forrester Research analyst, less than successful.
Principal analyst Boris Evelson said BI environments are so often plagued by myriad moving parts like data sourcing, data modeling and data cleansing components that must be properly assembled. To get the desired end-user format in a report or dashboard, “there is lots and lots of effort that needs to be done to put these types of applications together,” said Evelson, who spoke at a Webinar this week entitled Next-Generation BI.
The matrix of bits that make up a complex BI environment can mean that a simple change to one source system needed for one report will “cascade” and affect other systems necessary for other reports. “There is no such thing as a small tiny quick fix, quick change to a business intelligence environment,” said Evelson.
Successful BI, said Evelson, is more like an art than a science, where IT departments need best practices and a large amount of experience – rather than just a few successful instances under its belt – to understand things like business-led data governance and differences between user types.
“But,” said Evelson, “it’s not all doom and gloom.” He explained that while BI environments may remain largely complex, vendor offerings are making headway in the areas of automation, pervasiveness and unification.
In the area of automation, he said, vendors are working on applications with built-in expertise that provides hints to users as to what they potentially want to do next based on what their peers did in similar situations. This is useful because data recovery remains a time-consuming and laborious part of BI, although it’s easy for search engines, said Evelson.
Emerging Web 2.0 technologies, such as mashups, he said, are helping drive unification of internal sources like a customer relationship management system and external data sources like partner systems. Vendors are also unifying BI that is historical with BI that is predictive, given the combined benefit of both.
Evelson noted that while all these BI technologies are certainly not a panacea, the right tools and best practices can help organizations attain what he called enterprise optimization, or the need to be highly-optimized in addition to being competitive. Enterprise optimization is driven, he said, by the need for businesses to find novel ways to differentiate themselves in an increasingly commoditized world.
Evelson listed three “pillars” on which enterprise optimization lies:
1. Efficiency: While customer relationship management and enterprise resource planning applications no doubt make organizations more efficient, Evelson said the fact that they are popular makes it “a level playing field.” That’s why BI is crucial.
2. Agility: Most applications, especially the hard-coded ones with rigid data models, are not very agile, and “today’s world moves way too fast for a typical application to keep up with that,” said Evelson. But new architectures like service-oriented architecture (SOA) and Web services are allowing organizations to more easily piece together their infrastructures.
3. Effectiveness: Evelson called this the “last frontier for enterprise optimization,” saying it will help organizations make better decisions especially as BI has now moved to front office applications.
Another guest on the Webinar, Dale Skeen, chief technology officer with Sunnyvale, Calif.-based business process integration technology vendor Vitria Corp., said that next-generation BI technologies will allow organizations to act upon intelligence quickly, associate business context with meaning from derived data, relate internal and external data, and compare real time and historical data to make predictions.
Such innovation, said Skeen, is not to be found in traditional BI products, rather those that recognize emerging technologies like SOA, Web 2.0, software-as-a-service to name a few.
While BI is increasingly pervasive on desktops and integrates with different systems, there is certainly room for more BI to spread its presence, said Evelson, like “integrating search in BI, integrating e-mail in BI, integrating content management in BI, is something that is definitely happening currently and in the near future.”
Along the same vein, BI pervasiveness means users will want to access their applications remotely whether online or offline, and he said vendors are focusing on that area as well.