Focus On: Analyzing Your Data: Tool Choices Abound in BI Space

If the market for business intelligence tools isn’t booming, it’s at least beginning to ignite inside companies that have implemented data warehouses and data marts to help them keep track of products, services, customers, suppliers and other business-critical information.

These tools – which include decision support systems (DSS), query and reporting software and online analytical processing (OLAP) systems – are hitting their stride for a number of reasons. Processing power has become more affordable on the AS/400 and data warehouses are sprouting up across many industry segments. Meanwhile, there’s been a shift in the data analysis paradigm that’s made the esoterica of custom query writing and detailed report generation accessible to a broader range of users.

Most recently, however, the biggest factor helping to grow the business intelligence market has been the expanding World Wide Web, and the use of intranets and extranets to distribute data to internal and external users. Over the last six months, business intelligence software vendors have begun to Web-enable their analysis and reporting tools, or introduce new products that use HTTP as a transport protocol for carrying business data between data warehouses and users.

Broadening the Reach of Business Data

“We have been surveying tools providers and about six months ago half of them said they have plans to Web-enable their products,” says Henry Morris, program director for data warehousing and information access at the Framingham, Mass.-based International Data Corp. “It’s the economics of the situation: making the data available to many more users without requiring lots of training.

“The Web interface can emulate database searching patterns in a self-sufficient way, and browsers are making the information available to more people.”

According to IDC, the so-called Enterprise Decision Support market is looking pretty healthy. IDC places online analytical processing (OLAP), user query and reporting, and executive information systems (EIS) in this category. Morris says that in 1996, this was a $1.28 billion business that grew 21.4 percent to $1.56 billion in 1997. Some categories grew faster, he says. OLAP tools for multidimensional analysis, for example, was a $378 million market in 1996, and grew 36 percent to $515.6 million in 1997. The market for decision support system (DSS) tools built on these products grows in the 20 to 30 percent range each year, Morris says.

These tools have historically been used by a narrow, vertical slice of the company, usually IT staffers creating complex queries and reports.

Lately, however, software tool vendors have begun to target more users within the company with their products.

“Business intelligence tools now provide a breadth of reporting types and analysis for different roles and needs of users,” says Scott Lawrence, senior manager of strategic alliances at Cognos Inc. (Burlington, Mass.) “Users now range from analysts to managers to consumers, all of whom want access to data. We’re offloading the backlog of reports sitting in the IT department, because now, all these people can be report authors. It gives users self-service reporting.”

Cognos isn’t the only company eyeing the potential for bringing analysis and reporting capabilities to users at all levels of technical expertise. Alan Missroon, VP of marketing at CorVu Corp. (Roswell, Ga.), describes the range of users that should be able to participate in business data analysis.

“We realized early on that there would soon be a vast difference in company personnel who would use this kind of product,” says Missroon. “We saw a ‘true end-user,’ who creates reports with bar graphs maybe once per week, similar to users of executive information systems (EIS).

“Then moving up the complexity scale we see business analysts who are doing more ad hoc querying and statistical analysis. They are interacting with the data and calculating on the fly. Then at the upper end of the scale, there’s the hardcore IT user doing production reporting and report scheduling and creating complex queries.”

Of course, easy setup, configuration and management has been the mantra of IBM's AS/400 group all along, and the fact that business intelligence tools vendors are now tailoring their products to less sophisticated groups of users appears to make the tools an easy sell to AS/400 shops.

"The AS/400's existing market is helping BI tool sales grow," says Mark Wulf, business intelligence segment manager for AS/400 partners in development. "Since its inception, the platform has done very well in the non-technical business market space, and with our partners, we can apply that to the BI space, too. A key differentiator between us and the Unix boxes is the value proposition of the AS/400, the ease of use and integration. Our systems don't require as much tinkering as UNIX systems."

Wulf also points to the processing power available with the 64-bit architecture, which is now affordable to more mid-sized and smaller companies. And now the AS/400e Series offers up to 20 GB main memory, while V4R3 has the processing technology to work with up to 40 GB of main storage -- another key feature that makes data warehousing and business intelligence viable to more users.

All this horsepower and memory space simplifies data analysis for AS/400 by allowing users to combine the data warehouse itself and OLTP on the same box, continues Wulf.

"Our customers have found that when your operational environment is the same system as your data warehouse environment, a lot of the difficulties, such as getting the database information into the data warehouse, are greatly diminished. Difficulties arise when the OLTP is done on the AS/400 and the data warehouse was on the UNIX box. Our OS and database perspective made us ensure that there is one single box with multiprocessors, and to parallelize everything on multiple processors. And beginning with OS/400 V3R1 back in 1994, we let users create indexes faster and cleanse data faster."

Several data warehousing and business intelligence tools vendors have aggressively pursued AS/400 users by shipping products that have been optimized for the hardware architecture and DB2/400 database processing. The NGS-IQ integrated suite of business intelligence products from NewGeneration Software Inc. (Sacramento, Calif.), for example, is one of IBM's value-added enhancement (VAE) partners for AS/400 server models. That means the NewGeneration Software is authorized by IBM to integrate the AS/400 server model into a customer's client/server environment when the AS/400 is an integral part of the customer's data warehousing, online analytical processing and business intelligence system.

Taking a slightly different tack, Business Computer Design International Inc. (Hinsdale, Ill.) sells PC-based data analysis and reporting tools called Spool-Explorer/400 and E-Z Pickin's, aimed at simplifying the process of downloading, viewing, and manipulating AS/400 spool files. Users can manipulate the extracted data using a large library of functions, filters, and sorting features, as well as view the data in a variety of graphical formats.

The Web: Is it the Perfect Data Analysis Medium?

The benefits of putting your data analysis on the Web? Ease of management, low maintenance, rapid deployment, low cost -- the check-off items we’ve heard all along from Internet evangelists. The main goal among business intelligence software vendors, however, is to get information from a data warehouse or data mart to more people in the company, and to do it more efficiently than ever before.

From the looks of the most recent spate of business intelligence product announcements, most vendors haven’t overlooked the advantages of deploying critical business data over the Web. Some companies have already introduced Web browser clients and Web server modules. Usually, the Web server handles the connection to the data warehouse, which resides on an AS/400 or other server system. In between the browser client and data warehouse sits a Web server, often accompanied by a Windows NT or Unix application server.

"Network technology and the Web have really helped the data analysis market grow," says John Hughes, senior VP of sales and marketing at Silvon Software Inc. (Westmont, Ill.) "Web browsers are ubiquitous, they’re a common means of communicating, and you can access data on the Web from almost anywhere. It’s relatively easy for a large company to maintain and update Web-based information."

Silvon’s DataTracker is a data mart application that runs natively on the AS/400 and Windows NT.

Putting data analysis and reporting tools on the Web eliminates much of the setup and management work required by traditional client/server architecture. So not only is the Web delivering data to more users, it’s also making the chore of delivering that data easier for IT managers.

"In the old two-tier client/server architecture, to do the query processing and analysis, you have to put ODBC drivers on every machine in the company, and then you have to manage them," explains Nick Halsey, vice president of corporate marketing at Brio Technology (Palo Alto, Calif.)

"With a three-tier Web-server architecture, you can put an NT box next to the AS/400 and a put a Brio Server on the NT box and you have one thing to install and maintain. Then if you put the Web server on the NT box, all the clients can be managed from one server and you can use HTTP as a protocol to the clients. Now your business analysis is on your intranet. "

What’s more, continues Halsey, when Brio releases an update to the Brio Enterprise Server, because it is on the Web, every user will be automatically upgraded when they log on. This is also a cheaper way to add users than adding individual Windows clients. In addition, reports can be distributed electronically at any time, instead of printed out on paper.

"If I distribute to a three-tier architecture and the Web, I don’t have to print and then distribute reports, I can just set up the Brio server to e-mail them," Halsey says.

Indeed, as less sophisticated users access the information in the data warehouse via the Web, a new class of business intelligence user arises.

These people aren’t creating ad hoc queries or analyzing multidimensional data, they’re accessing information that has already been wrung through the IS department and is in the form of a report or other easily digestible form, says IDC’s Morris.

"When considering Web-deployment of business data, the question arises, ‘Are these people going to directly access the data warehouse?’," asks Morris. "Many people will only need access to the reports themselves. The increasing trend toward publishing enterprise reports on the Web server on the middle tier lets users browse through reports and pick out important parts and even allows for some publish-and-subscribe to customize reports for specific users."

For its part in the "Webification," if you will, of business analysis tools, Business Objects SA (San Jose, Calif.) offers WebIntelligence 2.0, the latest release of its enterprise query, reporting, and OLAP package for the Web. The latest version offers features intended for extranet deployments, a new OLAP module for drilling and charting, and new options for report creation and distribution. Business Objects has certified WebIntelligence 2.0 for DB2/400 version 3.7 when it is used with Windows 3.11, Windows 95, Windows NT 3.51, and Windows NT 4.0. Each certification uses IBM CAE as the database middleware, and means that these connectivities are supported for database query and management of the BusinessObjects relational repository.

Custom-delivered Data

MicroStrategy’s DSS Broadcaster is an example of a custom report and data analysis product. It’s a Windows NT-based product that works with the company’s DSS Web OLAP tool to deliver personalized business analyses to users based upon preset criteria, rather than requiring users to continuously query the data warehouse.

"Highly summarized data isn’t meaningful to each individual customer," says David Sherwood, VP of marketing at MicroStrategy (Vienna, Va.) "DSS Broadcaster allows you to create reports from data warehouse information that’s particular to each customer and provide it to them via e-mail, fax, pager or voicemail."

Sherwood explains that MicroStrategy can help users extend their data warehouse information outside the company as well. He says users can reach their supply chains, which can access a data warehouse via the Web or dial-up line to help manage inventory, or the data can be broadcast to suppliers using DSS Broadcaster. He adds that users can also keep track of customer loyalty, a big challenge in any business. Customer information can be relayed to suppliers and can help the company target specific customers with specific information from the data warehouse.

With the movement of business intelligence to the Web, not only are industry analysts seeing a new class of business data user, they’re identifying a new category of server software. The business intelligence market is moving to consolidating all analysis on a mid-tier server, and personalizing reports for specific recipients.

According to Forrester Research (Cambridge, Mass.), this is the introduction of the "analysis server," where the OLTP environment and Web server are on one box, and the Web handles data distribution. Then the application server acts as a business intelligence server.

"The reporting software market is collapsing into what we call interactive analysis," says Frank Gillette, an analyst at Forrester Research. "Vendors are providing access to all the features through a Web browser, so there is one portal for report analysis and data mining."

IBM's Wulf adds that the company is working to ensure that its data analysis tools partners not only work well with ERP applications from vendors such as Baan, PeopleSoft and SAP, but that they are also accessible to people who are only familiar with Web browsers or spreadsheets.

"While the query tools themselves might be easy to use, a lot of people are not familiar with the database aspects of the system, and we're seeing the progression of these tools aimed at non-database people. The query tool vendors incorporate metadata on the front end, but we're trying to have them reduce the number of database constructs on the front end. So on the query tools users just see a big button-press icon for creating reports or running analyses," says Wulf.

Clearly, the Web is helping make data warehouse contents accessible to a growing number of non-technical users, and it is changing the way users analyze business information to help them make critical strategic decisions. There may come a time when a company either invests in a widely deployable data analysis tool, or watches its competitors flourish because they beat them to it.

"Businesses are reaching a point where they have to use data analysis tools because their competitors do," says Brio’s Halsey. "They have to understand their customers and competitors by maintaining transactional data in a data warehouse or data mart. Even with all the focus on Y2K right now, there’s healthy growth in the business intelligence tools market, which indicates users are taking these products very seriously."