In-Depth

What Is Business Intelligence? Part I: The Analyst Assessment

These days, the notion of business intelligence (BI) seems very much in flux—perhaps because everyone wants a piece of the pie.

• By Stephen Swoyer
• 01/25/2006

Ask three people to define business intelligence (BI), and you’ll likely get three very different answers. Or perhaps you’ll get three answers that agree in general terms but still differ in important respects.

To a degree, that’s always been the case. But there’s something more at work here. It’s almost as if the very notion of BI qua BI is very much in flux these days—perhaps because everyone wants a piece of the BI pie.

Enterprise applications vendors, for example, see BI as a natural complement to the operational data that’s already sitting in their ERP systems. So why not use Oracle Business Intelligence or SAP NetWeaver (nee, Business Information Warehouse (BW)) to get at it? Relational database vendors, on the other hand, frame BI as an extension of the RDBMS itself—i.e., they already own the data, so why not use their tools to report against and analyze it? Meanwhile, business intelligence pure-players say BI is all about the platform; with a healthy dose of performance management thrown in for good measure, as a way of closing the proverbial loop.

This debate might not amount to a struggle for the very heart and soul of BI, but it’s confusing nonetheless. In the midst of this confusion, then, it’s worth getting back to the basics. What is BI? Or—more properly—what is its essence?

In the coming weeks, we’ll speak with analysts, vendors and users about the ever-evolving state of BI.

A BI Arms Race

In most respects, says BI analyst Mike Schiff, a principal with data warehousing consultancy MAS Strategies, BI is—in the words of Talking Heads braniac David Byrne—the same as it ever was.

“BI is a whole spectrum of technologies that consist of multiple complementary products, ranging from simple queries through OLAP analysis and enterprise reporting, all the way through data mining and predictive analytics as applied data mining. That’s the boilerplate [answer],” he comments.

There’s more to it than that, of course, Schiff concedes. “Much of the tugging that you see is really fighting over market share. The enterprise application vendors don’t really want to leave money on the table, so they want to cover both the operational aspects and the analytical. That’s why they position it as ease-of-introduction versus best of breed. The BI pure plays, they obviously want to protect their own market share, but they also want to keep growing, so that’s why you see them pushing their BI platforms and their business performance management (BPM) tools.”

Interestingly enough, these contradictory motivations actually have a positive impact, Schiff argues. To the extent that the ERP vendors, database vendors, BI pure plays, and smaller BI players innovate to protect their own market segments, they also determine the evolutionary direction of the BI space itself. Like our single-celled forebears (whom some believe were spawned in puddles of primordial goo), the BI tools market (not to mention our understanding of what is meant by business intelligence) is continuously evolving toward ever greater sophistication and complexity.

Like MAS Strategies’ Schiff, TDWI director Wayne Eckerson starts with BI boilerplate. “TDWI's definition [of BI] is the ‘the processes, tools, and technologies for turning data into information, and information into knowledge, and plans that lead to positive actions for the organization,’” he observes.

As far as Eckerson is concerned, however, BI is an altogether more intricate enterprise. “BI is like a data refinery. Like an oil refinery, it processes a raw material—in this case data—and turns it into a multiplicity of information products for consumption by the business. The business uses the information products—reports, analyses, data mining models, dashboards, scorecards—to test and refine their assumptions about what drives the business,” he explains.

To what end do we use and consume BI tools? According to Eckerson and other analysts, insight is the quintessence of the thing. “BI is the cornerstone of a learning organization, one that uses facts to validate intuitions and make steady progress towards achieving strategic objectives,” he argues.

BI isn’t a fixed and immovable practice, however. More to the point, the expectations of BI consumers tend to change (often dramatically) over time—typically in response to the evolution and maturation of BI tools and technologies.

In the 1980s, for example, the term “BI” was mostly used to describe the process of reporting—largely by means of legacy reporting tools. By the 1990s, “BI” more properly described the practice of analysis—using spreadsheets, OLAP tools, and query/reporting tools. In both cases, the expectations of users matured (more or less) in tandem with the evolution of then-cutting edge BI tools.

If anything, Eckerson suggests, the rate of BI evolution may actually be accelerating: over the last five years, he argues, “BI” has gone from a mostly desktop-based enterprise to a largely Web-based experience, capable of supporting not just core constituencies—i.e., knowledge workers, analysts, business decision-makers, and C-level executives—but also external customers or suppliers via the Web. Over the next five years, Eckerson suggests, our understanding of what is essential to BI will evolve still further, to include the use of dashboards, scorecards, planning tools, and BPM suites.

Some Still Grapple with the Term

Those of us involved with BI technologies and practices often take BI for granted. But the general public is still grappling with the term. Even mainstream business media seems largely in the dark. As one business editor for a major US daily newspaper said recently of BI, “So, it’s just a bunch of servers in a room somewhere?” It’s a bit more than that, of course. A few folks even mistakenly assume that business intelligence is a form of industrial espionage—commonly called Competitive Intelligence. Although the man on the street can’t define BI accurately, business people who rely on information technology usually can.

“Wide-spread adoption of the term ‘business intelligence’ is fairly recent,” says Philip Russom, TDWI’s senior manager of research and services. “In the ‘90s, most users would talk about ‘building a data warehouse’ or ‘writing a body of reports,’ whereas today many talk about ‘implementing a BI program.’ The shift in terminology reminds us that warehouses and reports are technology pieces that support the intelligence of the business. Speaking of support, I sometimes pine for the old term ‘decision support,’ which is a pretty good description of how BI technologies help a business.”

An Evolutionary Stutter-Step

The evolution of a concept is a tenuous thing. Many aggressive adopters have already embraced Eckerson’s forward-looking view, while others maintain notions of BI that are only slightly more advanced than the bread-and-butter reporting of the ‘80s or the spreadsheet-driven analysis of the ‘90s. “Companies are still evolving through the cycles at different rates,” says Eckerson. “Some are still stuck with desktop and production reporting tools and are willing to call that BI. Others are automating decisions and actions using highly integrated BI environments.”

MAS Strategies’ Schiff, for his part, thinks our understanding of BI will continue to evolve in a kind of stutter-step—with ambitious vendors leading the way and skeptical customers following, albeit slowly, in fits and starts, as BI innovations are demonstrably linked with the accretion of competitive advantage.

Call it an equilibrium—or an evolutionarily stable strategy—of sorts.

Schiff cites BI standardization as a case in point. “Right now, no vendor has the best product in every single category, contrary to what they might claim, so in that [respect] it’s a tough sell, yes—but there’s definitely a trend of users not wanting to play referee between all these different vendors. There’s already a trend toward ‘all things being equal, let’s try to get all we can from a smaller set of vendors. It may not be optimal, but it’s good enough,’” he comments. “This is a case where right now [the BI vendors] are leading the market, but in a few years, it’s going to be the conventional wisdom.”

To be continued...