By Todd Wasserman
By now, it's hardly unusual to find an organization that's implementing business intelligence (BI) technologies, but not every enterprise is equally intelligent in its execution of BI strategies.
One of the biggest mistakes often occurs in the IT department, which confuses BI with other business functions, like accounting, and addresses it primarily by buying new hardware and software. Another common gaffe is on the data-gathering end, in which poor data leads those analyzing the information to make inaccurate assumptions.
It's very important to first figure out a BI program and then buy the gear, according to Betsy Burton, an analyst with Gartner Group. "Otherwise, you're going to buy technology that you may or may not use," she says.
BI can be compared to military intelligence adapted for corporate ends. Specifically, BI entails the use and analysis of data that can help people working for an enterprise to discover, innovate, manage, or optimize it with the goal in mind of driving financial benefits to the organization. IT managers need to understand the types of data available and the types of data they already collect before buying BI tools -- software applications made by many top technology companies these days.
A good example of BI in action is using data gathered from customer relationship management tools to create a customer segmentation marketing strategy. A good BI system will let you quickly see patterns where you hadn't before and then act on them. For example, a music company might discover that customers aged 60 and older are buying rap CDs, perhaps as gifts for grandchildren or for their own enjoyment. That pattern might convince the company to start advertising in more senior-skewing media.
But to get this far, a company has to decide what it will be looking for, which requires a plan.
Developing a Business Intelligence Plan
Burton suggests the following steps to put together a plan that can help an organization execute a BI strategy:
- Identify the business objectives Figure out which metrics will be used to claim success. Boris Evelson, principal analyst for business intelligence with Forrester Research, says companies should have a "data dictionary" so that, say, finance and IT can agree on terminology
- Define how BI tools can help Look at the organization's business objectives and examine BI as a way to measure, monitor, or manage processes to help meet those goals.
- Compare this perspective Contrast BI's potential to currently employed tools and technologies and see how they stack up.
- Build a road map Determine where the organization needs to go and balance that strategic perspective against day-to-day requirements.
Burton proposes that CIOs form a Business Intelligence Competency Center (BICC) made up of staff members from different departments to oversee an organization's BI strategy. According to Gartner's research, about 40% of companies have a BICC, a number that has doubled in the past two years. BICCs can take many forms, as there's no hard and fast rule about how often they should meet.
There is one requirement all BICCs should adhere to, though, Burton says. The center should include some C-level execs like a chief operating officer, as well as people from the IT department and analysts.
Ensure High-Quality Data Is Gathered
After a company has assembled a BICC and identified its goals, the next step is to make sure the data being gathered is of high quality, says Marcia Kaufman, a partner with Hurwitz & Associates, a research and consulting firm.
"If the data isn't collected properly, it can't be analyzed properly," Kaufman says.
Another issue is the sheer volume of information gathered. Forrester's Evelson says that Capital One, the financial services firm, has about 40 pages of data on everyone in the U.S.: "If you don't have a meaningful strategy, you can sit there for years trying to organize the data," he says.
Luckily, businesses have more tools than ever to assure quality data collection. For instance, there is analysis of unstructured text, which applies to searches across all data. If an organization has a potential issue with a battery recall, for example, an unstructured text analysis might reveal that battery issues are popping up in different types of communications, ranging from call centers to customer relationship management programs.
"If you try to wait for this to show up in more traditional forms, it takes a long time to show up," says Fern Halper, another Hurwitz partner.
Collecting the right data might sound like elementary advice, but Halper cautions that it's often harder than it seems, particularly if a company has recently merged or acquired another company, as each uses different types of applications.
"It's a major challenge to develop cross-enterprise applications," Kaufman says.
The best advice is to figure out a BI strategy before buying the technology or gathering the data. Says Kaufman: "If you don't have a clear, consistent understanding of business processes, it's going to be really challenging to get useful information, regardless of the BI tool."
Todd Wasserman has more than 15 years' experience writing for The New York Times, The Industry Standard and Business 2.0, among other publications. He is currently news editor for Brandweek magazine.