What you don't know can hurt you: Business Intelligence was once a "big company thing," but...

As a term, "business intelligence" (BI) sounds like a meaningless

buzzword that could, with a bit of imagination, apply to almost any aspect of IT. Ironically, BI is all about precise definitions and accurate information. From a management accounting perspective, it addresses a real and pertinent issue; the deficiencies many organizations have when it comes to reporting financial and other explicit business information.

The problem organizations have with reporting is not that it isn't done, but that it isn't done efficiently, explains John Joynt, president of GSS Marketing, an organization that provides BI solutions to mid-market companies. "We find that companies are incurring costs for sales analysis, financial analysis, vendor analysis, contract analysis, all that stuff," he says. "Everybody does that. But they are, for the most part, not satisfied with what they've got."

Peter Carson has had a similar experience. Carson is president of Envision IT, a company that provides Microsoft Business Intelligence solutions to the mid-market. "We haven't gone into any organization that I can think of that's had good strong reporting coming out," he says.

What Carson finds is rampant in organizations is something he calls "Excel hell," where financial numbers are consolidated by a team of people who manually assemble data on Excel spreadsheets. Says Carson, "The biggest symptom is the time to get information. If you need six people to pool various numbers together that finally get consolidated in the big picture, that can be 30 to 45 days after end of period before you get a snapshot of what's going on in your organization. And if you do that on a monthly basis, you're always a month or a month and a half behind."

[ILLUSTRATION OMITTED]

Hitting the KPIs

According to Joynt, employees often find relationships are strained over incomplete reports, key numbers that don't balance, or disagreements about who is performing and who isn't. Joynt gives the following illustration: "I say the number is 2, you say it's a 3'. And you spend the whole meeting arguing about whether it's a 2 or a 3, and the right answer is 4."

The key to any business intelligence project is zeroing in on the organization's key performance indicators (KPIs), Joynt explains. These can pertain to sales, profit, manufacturing, HR, or any parameter that is deemed vital to the business. The issue is to find out what information needs to be captured to properly monitor the KPIs. Once that is accomplished, the technical aspects can be evaluated.

The grunt work in these projects results from the fact that much of the data that's stored in an organization's IT systems is typically not accessible in a form that can be easily mapped to a KPI. For example, an organization might have information about how many widgets it sold in 2004, but it might not be easy to find out how many of these were sold in Manitoba.

This occurs for a variety of reasons. The data collected might exist in disparate transactional databases that abide in different locations. The systems might have different data architectures, making them difficult to consolidate. The data in them might have redundant information that has to be removed, or as they say in IT terms, cleansed. In some cases, the data required might not be collected at all. But even in the best of circumstances, databases that support day-to-day business are optimized for quick and easy updating, but not to support the kind of broad and complex querying that business reporting requires.

The traditional BI solution involves the creation of an intermediary database, called a data warehouse, that is optimized for reporting. The data warehouse is populated with selective subsets of data that are deemed relevant. For example, a retail company may need to know the percentage of customers paying by credit card. The transactional database that has this information will also have credit card numbers and expiry dates. Unnecessary information wouldn't go into the data warehouse.

Better answers

Data warehousing has traditionally been a big ticket item. But with offerings targeting the mid-market, such as SQL-based Microsoft products, the entry barrier is dropping. Furthermore, the definition of what constitutes BI is broadening, and BI can be approached in a modular fashion. "From our point of view, business intelligence means any sort of information that you're getting out of your systems," explains Peter Carson. "That can be a digital dashboard that you go to on your portal site, it can be a Web-based report, or an e-mailed report that comes out to you, or it can be the more traditional cubes and analysis (typical of data warehousing) kind of thing. What we're seeing is that any decision maker in an organization has a need for business intelligence."

Fred Gillespie, GSS's VP of sales, expands on the more general use of BI within a company. "What that really means is that more people are able to make better decisions through the availability of better data. And Microsoft themselves talk about their whole strategy as giving the mass of the people inside a company--not just the executives--the kind of information they need to make better decisions every day. So instead of just the CEO making a good decision, there are key performance indicators for every department. For the mail department, there's a KPI. What determines success in the mail department? What Microsoft is really trying to do, and where we're going, and where I think the whole industry is going, is making reporting available to all the employees of a company. When a customer asks them a question, or when another employee asks them a question, they give a better answer than they would have given before."

Of course, the process of making management information widely available has to be carefully orchestrated. Indeed, this can uncover deep organizational divisions over roles, responsibilities, and ownership. In general, there tend to be two types of BI user. The so-called power user employs BI in the traditional sense of slicing and dicing aggregate numbers to provide detailed strategic reports to senior management. The new "class" of users consists of those who use BI to monitor factors affecting their own area, such as their sales performance compared to other regions.

Another force that appears to be driving BI is the need for regulatory compliance. Sarbanes-Oxley, for example, is very explicit in its requirement that an organization keep tabs on all factors that could reflect on their financial health. The application of BI in this case is clear.

However, in spite of the popularity of compliance solutions, the real driver behind BI continues to be the need to make better business decisions. As GSS's Gillespie points out, the more recent approach to BI reflects the reality that business decision making is not limited to the boardroom. "A well run business is not a business that made one good decision. A well run business is a business that makes a lot of good decisions every day."

Jacob Stoller is an independent writer and researcher based in Toronto. He can be reached at jacob@stollerstrategies.com.