Test Your IQBy Michael Fitzgerald | Posted 01-07-2007
Test Your IQ
It took four years for Andrew Woje-Wodka to figure out how to make business intelligence matter to the business side at Del Monte Foods Co. Wojewodka, director of business systems and decision support services at the $2.9 billion, San Francisco-based food giant, had been using software from Cognos Inc. to report on matters such as the fill-rate, an important measure of how the company was filling customer orders.
"We were giving them these pretty dashboards with fill-rate by product or geography or customer, but it was just wasn't being adopted well" by business users, Wojewodka says. He finally realized the dashboards weren't telling business users why fill-rates were what they were. There was no way to tell if a fill-rate target was missed due to bad forecasting, issues with Del Monte's distribution channels, or some other factor. So he started using analytics tools from Cognos to present potential causes. That got the business side's attention.
"It was like the light bulbs went off," he says. Now, Del Monte can't seem to get enough BI data. In the last year, Wojewodka's 24-person department has seen BI project requests jump from 109 to 131despite finishing 62 such projects and merging 13 others. BI projects are tracking, for example, whether customer service is improving, whether inventory stocking is meeting its targets, and whether top-line and bottom-line growth is where Del Monte wants it to be.
Del Monte isn't alone in its desire for business intelligence tools, software that helps companies track their performance and their relationship with customers. The Society for Information Management placed BI third as most important technology of 2006. And recent CIO Insight research found that CIOs said the two technologies that will make the most significant contribution to corporate strategies this year are business intelligence/data mining, and an offshoot of BI called business process management.
Gerard McCartney, acting CIO of Purdue University, says that business intelligence tools are on the rise because the software is maturing.
"A lot of dashboards in the 1990s were just kind of gimmicks," he says. But more important, BI is ascending because businesspeople finally understand it. "Executives have become a lot more sophisticated about data, and what it means, and I think for the first time they're really realizing that they have to have [BI]," he says.
The market numbers back up McCartney: IDC projects 10 percent annual growth for BI software sales through the rest of the decade, taking the market from $1.7 billion in sales in 2005 to $2.7 billion in 2010. IDC analyst Dan Vesset agrees that, at this point, every large company uses business intelligence software.
But as companies further embrace business intelligence, they will run into a host of challenges: ensuring data quality, defining how and what data should be measured, and putting BI tools into the hands of a wide range of users.
How and What to
How and What to Measure
Wojewodka believes the reason for the growth is simple: Once businesses start getting answers to their questions, they start asking more questions. "There's an insatiable appetite for more information and analytics," he says. "It feeds upon itself."
But as Del Monte found, just buying business intelligence tools doesn't mean your business will be smarter. Getting data to mean something useful to the business remains a challenge even now, more than 30 years after the seminal BI tool, the spreadsheet, hit the corporate desk. In part, that can be a problem of presentation, as Wojewodka found when he realized that his BI tools were telling people only what had happened in the business, but not why.
A more fundamental problem comes from the data itself, says Purdue's McCartney. "The dashboard is only as good as the data that's underneath it. If it's dodgy or ill-defined, when you get department-level people poking at data mines you end up with 'dueling reports,' " he cautions.
But getting well-defined data can be a minefield for CIOs. At Purdue, for instance, good data can help administrators by letting them respond when a class fills quickly, by setting up new sections. This could mean finding extra teachers and increasing a textbook order, tracking inflow and outflow of funds, and parsing the student body. But at Purdue, simply defining whether Nebraska is in the Midwest
can spark heated debate.
That's one reason the university took almost three years to plan its current $73 million SAP project, designed in part to give the university better reporting and analytics, for which it plans to begin by using the BI tools built into their new SAP system. Right now, most of the school's business intelligence is done with spreadsheets, via batch processes, or even with plain old pencil and paper. McCartney says these methods are simply too slow in the Internet era. "We're not just upgrading our software. We're looking at our entire business process," he says. Figuring out what needs to be measuredand how to measure itis a basic challenge for any organization that wants to have better insight into its business.
And, McCartney notes, it's important to get it right. "Let's be clear about this: Once the troops know the boss is looking at eight different measures, or 20 different measures, they'll be engineering stuff to maximize those measures, so they'd better be good ones."
But as BI becomes firmly established in corporations, CIOs can be faced with a new set of problems involving "systems creep."
BI tools often spread through an organization without much noise. Most come bundled with reporting or analytics tools, staples of business intelligenceIDC tracks six different categories of BI tools and 16 major application areas. "BI just comes with the stack, and there's no negative ROI impact not to use itit's just kind of there," says David O'Connell, a senior analyst at Nucleus Research in Wellesley, Mass.
While the market's largest players are pure-play companies such as Business Objects S.A., Hyperion Solutions Corp. and Cognos, no one company dominates the BI market. And many kinds of packaged applications have BI tools. CRM applications from Siebel (now part of Oracle Corp.), for instance, include robust business intelligence tools, and rivals have responded in kindto name just one, there are several dozen BI choices available on Salesforce
.com's AppExchange, a popular library of tools for the CRM package.
As users realize the kinds of questions they can answer with BI tools, they will often start using them on their own. This can be beneficial, in the sense that it can help people work more efficiently, but also potentially negative, in that numbers can be manipulated to skew what's really happening in a business unit. As Woje-wodka says, every department has its own metrics, and what Del Monte's Production Planning Inventory Control defines as acceptable levels of fill-rate may not make customers happy. So a business, instead of getting better intelligence, can run into the "dueling-reports" problem McCartney mentioned, as groups start using their departmental BI tools as weapons in business meetings.
Some CIOs have tried to mandate standard BI tools, usually based on a pure-play product such as SAS or Business Objects. That can help, to a point. But it's almost impossible to keep employees from using departmental tools, and it's probably not advisable, says Mitch Kramer, an analyst at the Patricia Seybold Group. Kramer says the analytic tools that come with a specific application are often better suited for working with that application than high-level enterprise tools, and businesses need to manage BI on both the operational and strategic level. The CIO can help by organizing rigorous definitions of what should be measured, and by which department.
As BI tools become more accessible in 2007, management challenges will undoubtedly get bigger, too. Vendors are making the tools more accessible, in part by making search tools better. They are including their own search features or adding software from companies such as Google Inc. and Convera Corp., a 20-year old provider of enterprise-oriented search tools, says Mark Smith, CEO of Ventana Research in Menlo Park, Calif. This is to spare BI users from having to know Online Analytical Processing (OLAP) techniques that search through different dimensions of database information, such as product, time of day, place and action (for instance, sale of a unit or shipping of a unit). In addition, better search will continue to be a market driver in 2007, as will the addition of features such as function templates, -preconfigured queries that save users from having to build their own.
Also on tap are ways to make business intelligence tools work better on mobile devices, such as the BlackBerry. This could help extend the BI-user market beyond the business analysts and managers who are its primary users now, and should mean that salespeople, in particular, will have faster access to information they need to sell clients.
And while BI tools typically are used to look at a company's own data, service-oriented architecture, a way of making it easier for applications to exchange data, means that firms can start to get data from third-party suppliers as well, which should help improve business operations.
All of these developments will challenge CIOs, but they also reflect the essence of what a CIO is supposed to do. "If the CIO is going to be a strategic partner in the business, he or she is going to be bringing tools that enable other people to do their jobs in a better and more strategic way," says Purdue's McCartney. "People at companies are smart, but they don't have the information now. BI offers the promise of having a better understanding of what's going on in your industry."