BI for the masses: user-friendly approaches
By Judith Lamont, KMWorld senior writer
“The business intelligence industry has been starved for user friendliness,” claims Sanjay Poonen, VP of worldwide marketing at Informatica. If that’s the case, a bevy of vendors are determined to set things right. Using a variety of approaches, software producers are finding ways to bring business analytics within the reach of knowledge workers throughout the enterprise. In addition, customers and partners beyond the enterprise are now also beginning to benefit from easier-to-use BI tools.
Typically, BI tools have been the domain of power users and analysts who prepared reports for business-oriented staff. Those feature-rich analytical solutions have allowed enterprises to make the most of their data warehouses. However, as organizations become less hierarchical and decision-making is pushed out into the enterprise, wider access to corporate data is necessary.
Informatica’s approach was to completely revamp the user interface of its PowerAnalyzer tool for “extreme usability,” as Poonen terms it. The design team was structured like a movie production crew, with a director and various personas representing different stakeholders, such as an executive or a shop floor worker. The resulting version, PowerAnalyzer 4.0, has a dashboard-style interface with alerts geared toward the user’s role. The dashboard might include a mix of bar charts, dials and small tables summarizing key information.
A guided set of steps allows users to seek out further details about the information they are viewing from the dashboard. For example, a user could move from looking at an overview of spending on the six top commodities to details on the number one commodity. From there, he or she could look at the top six suppliers, and then compare the performance of the top two in order to select a single supplier for the future. All of those steps would be accomplished by point-and-click or menu selection.
“Every knowledge worker should have information at their fingertips,” says Poonen. “This is the trend we see.”
When interactive capability beyond the enterprise is a requirement, ease of use is even more important.
“Customers and partners don’t want to go through training” says Andy Coutts, CEO of Databeacon, “yet they still want to be able to analyze the data themselves.” For example, GHS Data Management, which provides online prescription processing for pharmacy benefits, wanted to make its data available to its customers. Maine Medicaid and a group of self-insured companies supported by GHS need ongoing analyses of pharmaceutical sales to monitor costs and detect fraud and abuse.
“Databeacon was exactly what we were looking for,” says Jason Skeffington, data warehouse manager at GHS. “The documentation was thorough, and the front end was both easy to implement and user-friendly.” Unlike some BI tools, Databeacon’s Web-enabled functionality is identical for both intranet and extranet usage. GHS provided its customers with a live demo of Databeacon operating against the data warehouse, and “they dug right in,” recalls Skeffington. “They were able to come up with the numbers in a matter of seconds.”
Part of Databeacon’s speed comes from the fact that it creates compressed, aggregated OLAP cubes and delivers them to the information consumer. That function is performed by Databeacon’s publishing tools. Databeacon also automatically downloads the Insight viewer, a 700K Java applet that allows the user to analyze the data interactively and graphically. Therefore, all analysis by the user is being done locally. In fact, users can go offline and continue their analyses if they wish. Keeping analytical work away from the server also makes the system scalable to large numbers of users without the need for complex load balancing or extending the infrastructure.
“Going out to a heterogeneous user group means that the analytical tool should be adaptable to different skill levels,” Coutts points out. “With Databeacon, features can be turned on or off depending upon the user’s needs.” For example, skill levels can be set to novice, standard and advanced. On the horizon for Databeacon is a new feature, collaboration. As users interact with the data, they can notify other individuals, and both interact with a common set of data.
Because the most commonly used BI tool is Excel, another strategy being used to extend analytical capabilities through the enterprise is to deliver reports in spreadsheet format. According to Actuate, more than half of users already put the reports they receive from BI systems into the familiar spreadsheet format so they can more easily manipulate the data. Actuate removes the need for this step with e.Spreadsheet, which delivers, through a browser, data in Excel format from the Actuate iServer. The data is delivered as fully functional spreadsheets, including formulas used to calculate values and “live” graphs to support what-if analysis.
Single version of the truth
CoreData, which provides ATM transaction processing services to financial institutions and retail stores, is using Actuate's e.Spreadsheet solution to deliver reports to site supervisors, managers and ATM owners at more than a thousand locations worldwide.
“We deliver 50 standard reports,” says Cy Comer, CoreData’s administrator for Actuate, “and the customers can then set date ranges or other parameters for their analyses.” This process is much easier for the customer than the previous one, when they received a large, flat file and used client-side software to produce reports. Customers can also opt to receive reports in DHTML or PDF format.
“Often, people who need business intelligence do not have access to enterprise BI tools or may not know how to use them,” says Mike Thoma, VP of product marketing at Actuate, “so they tend to build renegade data warehouses and then use Excel for analysis.” The problem with that approach, he notes, is that different departments may not have the same baseline data.
“CEOs want a single version of the truth,” he adds, “and relying on data from a central server, rather than multiple data marts, accomplishes that goal. Actuate eliminates dueling spreadsheets.” Earlier versions of the report design product, e.Spreadsheet Designer, required IT-level skills, a new version released in March 2003 allows users to design their own reports by building templates using the spreadsheet paradigm and publishing them to the Actuate iServer.
DDB Worldwide, one of the largest advertising agencies worldwide, is using business intelligence tools from SAS to analyze data from an extensive study of consumer lifestyles and brand preferences. An 800-question interview of 21,000 consumers in 23 countries has produced 17 million pieces of information that must be organized and distributed to advertising and planning staff. Widely acknowledged for its leadership in creativity, the staff is not numbers-oriented, yet needed to understand the data emanating from the study. The data is stored on a server in Chicago, and the Web-enabled system is accessible to DDB staff throughout the world.
“The SAS system has had an extraordinary impact on our firm,” says Ned Anschuetz, VP of DDB. “Our team has access to a common data repository and a consistent way of interpreting it.” Through drop-down menus, graphs and tables, the SAS solution has enabled offices throughout the world to understand information they would not otherwise have been able to analyze readily. Users can order company-specific reports, compare a single brand across multiple countries, or compare several companies offering similar services. “The purpose of the study is to help plan ad campaigns and help clients plan product development,” explains Anschuetz. “We did not want to spend time learning to use a BI tool.” Ease of use has helped DDB team members focus on the content of the research rather than on the assessment process.
A strategy employed by SAS to make its BI solutions more effective is to develop a segmented set of interfaces targeted to different types of users. This allows users anywhere in the enterprise to contribute to the intelligence value chain (IVC), the term SAS uses for the process of collecting, storing and analyzing key data to achieve business objectives. Without this kind of differentiation,” says Don Hatcher, VP of Technology Strategy at SAS, “you end up with an elaborate product that is not particularly useful to anyone.” Domain experts, he observes, are not hired for their software skills, yet the ability to use data to make decisions at every level is critical to the evolving “agile enterprise.” True incremental business benefits, he believes, will be seen when organizations have an end-to-end IVC that allows full participation.
Beyond data warehouses
Not every important piece of data meant for business intelligence analysis fits neatly into a data warehouse. Some bits of information are too small or do not match the data structure in the warehouse. Other information may be contained in legacy systems that will never be incorporated into the data warehouse because doing so is too expensive.
“IT is inundated with the need to manage data that comes from many disparate sources,” says Jim Kanzler, CEO of Meta5, “but putting it all together in a data warehouse is not feasible or done in a timely manner. Like it or not, users are taking up the slack by integrating information dynamically at their workstations by cutting and pasting or developing customized integration macros.”
Business intelligence integration (BII) addresses this issue by integrating data at the application level, across multiple sources in an auditable way. That approach is particularly well suited for ad hoc queries and reports that require data from the warehouse, as well for queries that fall outside the scope of the data warehouse. Meta5 has a user-friendly front end that can deliver those results to the Web, Lotus Notes or a PDA. Meta5’s Tools or icons can be configured to point to various data sources or applications. By connecting those icons together, users can construct new applications that automate the cutting and pasting typically found in multidata, multiapplication environments. This approach is innovative in being able to deliver quick results in an auditable manner across multiple applications.
“Performance of business intelligence systems is not being evaluated correctly,” contends Kanzler. “Instead of measuring the speed of a query and response from the data warehouse, we should measure the time it takes from when a businessperson develops a question to the point at which he or she gets the answer.” Kanzler favors the deployment of data warehouses for business intelligence, but encourages users to be realistic about the limitations, and to seek out ways of including a broad range of inputs into their analyses.
Turbocharging BI systems
Appfluent Accelerator provides offloads reporting traffic from production systems to give users access to real-time data for decision support without negatively impacting enterprise system performance. It alleviates the bottleneck that results from having a single infrastructure handle operational, transactional and reporting functions.
HyperRoll is a database acceleration engine that integrates transparently into an enterprise’s existing database or data warehouse environment. Its aggregation and calculation caching engine optimizes access and analysis of summary information, producing significantly faster query responses.
Judith Lamont is a research analyst with Zentek Corp., e-mail firstname.lastname@example.org.