If your goal is to improve decision support through the use of analytic applications, put these highly valued features at the top of your wish list as you survey the marketplace.

Some Analytic Application Vendors

SoftwareVendor

Web site

Comshare Inc.

www.comshare.com

Hyperion Solutions Corp.

www.hyperion.com

MineShare Inc.

www.mineshare.com

SAS Institute Inc.

www.sas.com

Seagate Software

www.seagatesoftware.com

Walker Interactive Systems Inc.

www.walker.com

"Analytic applications" is the latest buzzword invented by industry analysts to describe the burgeoning market of online analytical processing (OLAP), business intelligence, reporting and performance measurement applications. Vendors such as Hyperion Solutions Corp., Seagate Software and Walker Interactive Systems Inc. are repositioning themselves as purveyors of analytic applications, and new startups such as MineShare Inc. are coming to market with products specifically designed as generic analytic applications.

Enterprise resource planning (ERP) vendors are climbing on the analytics bandwagon too, with SAP, PeopleSoft Inc., Oracle Corp., Lawson Software and others announcing products that deliver integrated analytic capabilities or partnering with vendors such as Hyperion and Walker to offer add-ons.

"Analytics" is not just another name for OLAP, which is one part of what an analytic application should supply (see A Typical Analytic Application Architecture). The goal of analytic applications is to furnish a wide range of functionality that typically includes:

  • data extraction and transformation tools to manage data sources
  • metadata creation and management for user-friendly data access
  • special "data cube" structures for use as temporary data storage areas
  • OLAP front-end applications for ad-hoc information analysis
  • production reporting tools to create, schedule and publish formatted reports
  • applets for accessing data from a desktop worksheet and/or a Web browser
  • information consoles for viewing data using tables, maps and charts
  • software agents for generating alerts and other event-driven actions

But what capabilities should you expect of an analytic application? The following 10 capabilities are key:

1. Leverage existing infrastructure

To earn an expedient return on investment, analytic applications must let companies leverage their existing network architecture, electronic mail system, accounting application data models and corporate databases. The analytics application should function "out-of-the-box" against application data with minimal additional infrastructure requirements, such as new database engines or data-access middleware. If the client applications used by the analytics application to display and analyze data can operate within a familiar worksheet product or Web browser, learning the analytic application on user desktops is easier.

2. Provide dimensional views of the business.

Analytics were not intended to process transactions, but to create information views that focus on such business-critical factors as customer, product, territory, salesperson or production line. Analytic applications should give users the ability to categorize business dimensions, a means to homogenize the coding of these dimensions across lines of business or locations, and the option to create user-friendly and meaningful names for the data and data views, a process known as "metadata."

3. Deliver cross-functional views.

Although analytic applications may be applied to single information domains, such as sales or production, they yield greater value when they consolidate information from across multiple functional domains. The analytic application must be able to extract and consolidate information from multiple application modules, not just a general ledger, or from heterogeneous applications and databases, then apply aggregation and transformation rules that streamline the data for analytical users. To generate useful information, sophisticated formulas may need to be applied to data, which may include financial and statistical data.

4. Offer an end-to-end deliverable.

Analytic applications are not pieces of a decision-support solution; they provide an end-to-end solution that may not depend on availability of a separate datamart or data warehouse. The analytic application may consist of a set of modules or components designed to work together and manage the complete decision-support workflow, from data extraction to data access using a Web browser. Analytic applications are supposed to be complete and integrated, more closely resembling an ERP suite than a multi-vendor, best-of-breed approach to decision support.

5. Deliver information proactively.

Proactive decision support relies on analytic applications that are event-aware and rule-driven; that is, able to recognize when certain events have taken place and then respond to those events by applying codified rules and triggering specific actions. An event may reach a specific time or date which prompts production of a specific report. Or an event could be fiscal, such as closing a period, which activates a period-end consolidation workflow to extract and aggregate production, sales and financial data. An event could also include reaching a threshold, such as exceeding a budget, that generates an exception alert for communication to a range of information consumers.

6. Allow for accessible information delivery.

Analytic applications must provide alternate ways for users to access information views. Some staff members may be trained to use an application-specific user interface to obtain information, but for others, this level of training may be unwarranted or too costly. Analytic application users should be able to obtain data views using tools they already have, such as a desktop worksheet or a Web browser, ensuring a wide range of users.

7. Offer enterprise scalability.

Analytic applications generally depend on widespread use throughout an organization to achieve optimum return on investment, so the application must have built-in scalability to handle large numbers of concurrent users, rapidly growing data sets and high demand for specific, popular data queries. The application architecture must be able to balance user loads across multiple data servers; use a database for storing its own internal persistent and temporary data that is itself scalable; and include features for optimizing the way that certain decision-support queries are created and run to make them as efficient as possible.

8. Recognize user roles.

The ability to set up user roles, which limit access to information based on an employee’s position, ensures security and adds an element of personalization to analytic applications. Roles such as manager, analyst or executive, for example, determine what level of information is accessible to individual employees and in what format, or which event workflows are appropriate for an employee to participate in. A manager may be restricted to a personal information domain, while an analyst may be allowed to roam across multiple domains. An analyst could have access to highly manipulable information views with extensive drilldown and on-the-fly charting, while an executive may only see specific performance metrics displayed in the form of a personal information dashboard.

9. Support multiple investigation needs.

Analytic applications must include a means to reach back into the transaction systems to complete drilldown paths that require source-level detail or to branch off to find other information in response to particular information investigation needs. Similarly, the analytic application should provide a variety of information export capabilities, especially to worksheet formats, HTML (hypertext markup language) pages or desktop database tables so that information views can be captured and used as snapshots for different kinds of analysis or formatting.

10. Provide application awareness.

Nobody wants to start from scratch building information views and reports in any analytic application when so much is known about the data stored in popular applications, such as the leading ERP systems. Analytic applications should offer some built-in intelligence and intra- or inter-module information views for popular ERP systems so that the analytic power can be used out-of-the-box or the pre-built views can be used as head-start templates that information systems or power-users can build upon.

If you keep these 10 must-have capabilities at the forefront when evaluating analytic applications for your business, you may surprise some vendors with your insight. And rest assured that your purchase will likely enhance the finance department’s decision-support capabilities.