Unlike many corporate executives, Jan-Willem Beldman says he doesn’t have a problem with the vision thing. A team leader at Mentor Graphics Corp., a computer hardware and software design company, Beldman can quickly pinpoint up-to-the-minute financial forecasts, peruse regional sales results, or view employee-performance statistics. And he does all this with just a few taps on a keyboard.
Beldman has lots of company. A growing number of corporate managers are turning to business-intelligence (BI) programs to help make sense of the reams of information produced by their ERP (enterprise resource planning), CRM (customer relationship management), and planning applications. According to one survey, about a third of corporations now use a dashboard. “Traditionally, ERP systems have been really bad at reporting,” Beldman explains. “They’re great transaction systems, but to get really flexible reporting out of one, you have to mix it with information from other sources.”
The BI system at Wilsonville, Oregon-based Mentor is typical of many now on the market. Called Essbase (from Hyperion Software), the multidimensional database engine funnels numbers from the company’s key business applications into several data cubes. Managers are then able to pull up specific information from them, thereby gaining an aerial view of what’s going on at the company. Just as important, the process takes only seconds. Says Beldman: “It’s analysis at the speed of thought.”
That’s been the lure since the phrase business intelligence first came into vogue in the mid-1990s. But the early promise of BI software far outstripped reality. Cobbling together ad hoc queries usually required the programming chops of Linus Thorvald — and the patience of Linus Van Pelt. Some later programs skirted that problem by prepopulating screens with a limited — and limiting — number of key performance indicators (KPIs). Even then, the bare-bones reports were usually only available at specified times, often weeks apart. That hardly qualifies as real-time reporting. But with recent gains in data archiving — and with much faster data retrieval — BI products are now spitting out up-to-the minute reports and nearly instantaneous responses to queries.
It’s about time, says Guillermo Kopp, vice president of cross-industries research at TowerGroup, a Needham, Massachusetts-based financial technology research firm. “Nobody is happy with end-of-the-day, end-of-the-month reporting any more.”
Release the Hounds
This general sense of unhappiness has been fueled in large part by advances in Internet searches. In an era when Google or Yahoo can rapidly and reliably find all manner of arcane information scattered across cyberspace, users have come to expect the same of BI products. And, in fact, it’s somewhat surprising that it has taken BI vendors so long to catch up with their consumer-driven competitors. Experts say the tardiness can mostly be attributed to the nature of search technology and the complicated — and diverse — structure of business data.
Indeed, Google’s recently released Desktop Search for Enterprise, which enables a business user to find files, E-mail, and pictures stored on a personal computer, isn’t suitable for sophisticated financial searches. Any mainstream keyword search engine lacks the guides and pointers — the intelligence, if you will — to accurately interpret complex queries aimed at unearthing relevant transaction data. What’s more, the new breed of hard-drive hounds — including Google’s Desktop Search and Microsoft’s MSN Search Toolbar with Windows Desktop Search — do not produce comparative sets of figures or generate graphical representations of key performance markers.
That leaves the field wide open to sellers of financial software, at least for the time being. ERP application vendors, which have been steadily encroaching on the BI turf, have been ramping up their BI offerings of late. This past May, ERP-giant SAP announced that it had completed a project with Intel and Hewlett-Packard aimed at improving the performance and flexibility of its Netweaver Business Intelligence application. Soon after, Oracle reported that it was working with Noetix Corp. to improve the integration of Noetix’s applications with Oracle’s Daily Business Intelligence. With Noetix, Oracle users are able to add additional KPIs to preconfigured dashboards and reports.
BI specialists, no doubt thrilled to see SAP and Oracle parking in their spaces, have been rolling out new products just as fast. SAS Inc. began shipping an enhanced version of its Enterprise BI Server in April. Five months later, Cognos Inc. came out with Cognos 8 Business Intelligence — a Web-services-based BI product. The Cognos launch came on the heels of rival Hyperion’s announcement that its core BI offerings would be available on HP’s 64-bit Itanium servers — a move aimed at dramatically increasing the scalability of Hyperion’s BI software.
Perhaps the most interesting development, however, came early in the year. In January, market-share leader Business Objects released the latest version of its much-trumpeted BI platform, BusinessObjects XI. The big news here: version XI features much tighter integration between the BI program and Microsoft’s ubiquitous desktop suite, Office. That should make it easier for workers to move between universes and, therefore, easier for them to analyze information existing in the Windows world. Says Jim Tolonen, CFO at Business Objects, in San Jose, California, “We’re doing everything we can to give our customers the ability to find answers, on the spot, to their important business questions.”
Customers seem to be responding. According to technology research firm IDC, in Framingham, Massachusetts, worldwide sales for BI software topped $3.5 billion last year. That’s an 11.5 percent increase in BI spending from 2003. The investments paid off, too. Says Dan Vesset, IDC’s analytics and data-warehousing research director: “Median ROI on a business-analytics project [in 2004] was about 110 percent.”
Despite the claim, many business buyers remain skeptical about the return on BI investments. In a 2004 market survey (conducted by Riley Research Associates for Noetix), nearly 40 percent of the respondents said the biggest liability of a corporate dashboard is the ROI. Another 30 percent of the surveyed executives said they were concerned about lengthy implementation times. That’s predictable, since early BI rollouts — particularly executive dashboards — often moved at a glacial pace.
The Riley/Noetix survey did offer one surprise. Only 14 percent of the executives said the most significant drawback to a dashboard was ease of use. That speaks volumes about just how far some BI tools have come. Older BI systems — including a fair number that are still on the market — require users to conduct complex searches with advanced techniques like Boolean logic. Now, says Kathy Quirk, a manager at Nucleus Research, a technology research firm in Wellesley, Massachusetts, both ERP-application makers and pure-play BI vendors “are trying to make their tools more accessible to a wider base of business users.”
Dashboards appear to be the preferred method for creating that accessibility. The desktop tool provides a familiar, centralized interface that nontechnical users can employ to navigate through mountains of enterprise data. Typically, a dashboard connects to a company’s far-flung transactional data systems — including ERP and CRM programs — via a master BI database.
IBM, a company that has staked much of its reputation and future strategy on enterprise information integration, sees plenty of potential in BI dashboards. In late 2004, the company teamed with Informatica, a Redwood City, California-based BI software developer, to produce a dashboard. The product, lyrically dubbed the Informatica and IBM Dashboard Engine Appliance, is designed for financial-analysis, risk-management, and regulatory-compliance applications.
In fact, it’s nearly impossible to find a financial-software maker these days that doesn’t offer a dashboard of one stripe or another. But not all finance managers are convinced of the merits of their offerings. “We’ve toyed with dashboards,” notes Steve Snodgrass, CFO at Graniterock, a construction and construction-products company in Watsonville, California, “but it may be more than our user community can really handle.”
Instead, Graniterock continues to use Business Objects’s Crystal Enterprise, a conventional BI reporting tool that the company deployed (on a limited basis) in 1999. “I’ve got about three or four people who centrally create reports,” Snodgrass explains. “If you let people create their own reports, it’s very hard to exert some level of quality control.”
Leery of Query
New releases of BI software — which are easier to administer — may ease concerns about dueling metrics. Nevertheless, a great deal of the corporate skepticism about current BI tools stems from problems with earlier BI tools, including executive dashboards and online analytical processing (OLAP) programs.
Interestingly, industry watchers note that OLAP software — once the bane of finance-department staffers — is staging something of a comeback. OLAP programs enable users to analyze information that has been summarized into multidimensional views and hierarchies.
“With OLAP, you are enabling people to slice and dice data and look at information,” says Kurt Schlegel, a research director at Stamford, Connecticut-based IT consultancy Gartner.
For all of their potential promise, OLAP products have long been hindered by slow response times and maddening interfaces. That’s changing. The latest multidimensional (MOLAP) products organize data in cubes — cubes that provide data views that can be rotated by the user. “Besides fast query response,” says Schlegel, “MOLAP solutions also provide the ability to do advanced calculations — for example, allocating costs to various dimensions to calculate profitability.”
Despite the advances, OLAP software products, like other BI tools, remain somewhat hamstrung by the sheer variety of data they must analyze. Metadata — and in particular data that is saved in the extensible business reporting language (XBRL) file format — could go a long way in easing that problem. Industry observers believe XBRL-enabled transactional systems will start making their way into mainstream use in the next five years.
Until then, Mentor Graphics’s Beldman stresses the importance of uniting disparate enterprise databases before splashing out on expensive BI tools. “It’s a difficult task to combine information from, say, your general ledger with your order system, because they don’t necessarily have a lot of things in common,” he warns. That’s why Mentor Graphics has worked with a number of its business units over the past year to define common data definitions. The goal, Beldman notes, is to make it easy for users to find the exact information they want, whenever they need it. “Design with that goal in mind,” he says. Then, no doubt harkening back to the early days of BI deployments gone wild, “build in small steps.”
John Edwards is the author of The Geeks of War.
The 800-Pound Gorilla Stirs
The rising corporate interest in business-intelligence (BI) software hasn’t escaped the notice of the world’s largest maker of corporate software. In May, Microsoft unveiled its long-awaited Office Business Scorecard Manager. Originally code-named Maestro, the software allows users to define key performance indicators and then view reports and charts that track those KPIs over time. The reports can be generated from a variety of databases or other enterprise-data sources.
In a way, Business Scorecard Manager is something of a throwback to the old days of formal BI reports. The software does enable users to forecast trends in real time, however. Thus, reports and charts can be called up whenever an employee needs to see a current view of key metrics. “Business Scorecard Manager is specifically designed for tracking a consistent set of metrics over time,” notes Rob Helm, research director of Directions on Microsoft, a Kirkland, Washington-based firm that tracks new products coming from Microsoft. “You might create on-the-fly views of those metrics — slice them up in different ways.”
Of even greater interest, Microsoft is also said to be working on bringing improved search/analytical capabilities to what is probably the most widely used BI tool: Microsoft Excel. Says one IT-industry analyst: “Everybody has spreadsheets that they’ve developed over time to track different types of business data.” Jury-rigging is hardly ideal, which is one reason Microsoft is expected to improve the search capabilities of Excel. Predicts Helm: “It’s highly likely that the next version of Microsoft’s spreadsheet program will have special features for defining and viewing information.” — J.E.