Not just any query-and-reporting tool will do. You have to take IS and end user needs into account
By Eric Rosenfeld
Turning data into useful information is one of IS' most critical roles these days, and there's no shortage of tools and strategies for doing it. The gamut runs from datamining tools that help you figure out what questions you should ask to production reporting tools that bang out high-volume reports from your production database.
Between these extremes lie two important categories. Managed Reporting Environments (MREs) let IS structure and filter data to turn it into information during the report development process. Managed Query Environments (MQEs) let the users create information when they need it.
The two approaches address a broad continuum of information delivery needs. At one end, the emphasis is on efficient distribution and careful control, and this is where MREs shine. The other end of the spectrum emphasizes ad hoc analysis of information, the usual province of MQE tools. To fill a given information need, you must take a close look at the trade-offs.
The Actuate Reporting System from Actuate Software concentrates on the MRE approach with its new Virtual Report Distribution (VRD) technology. VRD breaks a report into elements and delivers these report elements to the user. The report itself is virtual--it is physically located on the report server, yet it is accessible to the user as if it were local. Report pages are delivered to the user's PC only as needed, so VRD dramatically decreases network traffic.
Actuate reports act more like applications or WWW pages than like classic reports. A sales report, for example, may start with a table of contents that lists each region. The user can click on a region in the table of contents to see the region's summary data. Clicking on the summary data can produce salesperson data by customer for the region.
Actuate's report server holds pre-extracted, preformatted information --reports--so each user can see a report without having to rerun it. The report server also acts as the quarterback of the reporting environment. It handles data access, produces live reports, defines security, and stores report definitions. It can deliver a report to an Actuate client viewer on a LAN or WAN or to a browser via an intranet or the Internet.
MQE delivers flexibility
But what about users who never know what information they will need today? Managed Reporting Environments by their nature present a preordained path through the company's data. MQEs, on the other hand, let users decide how to combine data to address a business need. For this reason, MQEs typically require significant user training: The users must learn how to form queries and format and interpret the results. While this may not be appropriate for the entire rank and file, the flexibility that an MQE tool offers is often just what data analysts and key decision makers need.
BusinessObjects, an MQE tool from San Jose-based Business Objects Americas, delivers data--not reports and report fragments--to a user. Using BusinessObjects, IS breaks down a database into a hierarchy of objects, mapping columns, relationships, and computations from the database to business terms with which the user is familiar. The set of objects form an application--a "universe" in BusinessObjects-speak--that is distributed to users so they don't have to understand the complexities of the database itself. Working within their designated universe, users combine business objects to create queries. A user working in the Sales universe, for example, might pull subobjects from the Customer, Sales Rep, and Invoice objects to create a special-purpose sales report.
In contrast to an MRE's fixed report elements, an MQE places few bounds on the ways in which users can combine data elements. This freedom places a heavier demand on the user's insight and creativity, but it also delivers flexibility and decreases the user's reliance on IS for access to information.
Data access tools and vendorsCompany/Product |
MRE and MQE can and probably should exist together in any organization. But it is of paramount importance to apply the appropriate approach to each problem. A user with relatively little need for analysis may be frustrated by MQE--the effort involved in creating reports would seem too high. An analyst, on the other hand, who needs a flexible view of the data could find the boundaries of the MRE approach confining. Here are a few guidelines to help you choose wisely:
Look at data volumes. MREs are better at dealing with large amounts of data. IS can tune reports to optimize database performance. Tables of contents and other navigation features make finding information in large reports easy. And because MREs extract data for a given report periodically rather than on demand, delivering a large MRE report to many users creates a relatively small load on the database.
Look at output formats. Both MREs and MQEs present data well, but MREs can produce far more complex layouts of information. Actuate, for example, lets IS place data from multiple queries on a page, use different page layouts for each page of a report, or combine multiple reports on a single page.
MQEs, like BusinessObjects, don't require IS to be involved with each report. Users can produce tabular or matrix reports and even simple user interfaces for changing the query and the report format at runtime.
Look at report delivery. MRE report servers not only contain the queried information but also deliver it in pieces as needed. MQE systems always run against "fresh" data and allow for nearly limitless ad hoc tailoring of the queries to the user's needs.
Look at data quality. MREs make it easy to verify the quality of the information they deliver since reports can be validated by IS as they are built. The data extracted by an MQE must be treated carefully. Someone must determine that the query was devised correctly and that it really asks the question that the user thinks it asks. MQE products like Software AG's Esperant use artificial intelligence techniques to address this issue, preventing the user from asking questions that would return erroneous or misleading results.
With careful attention to these trade-offs, you should be able to find a good match between users' information needs and an MRE or MQE. //