Not long ago, companies built OLTP applications from scratch using COBOL or other 3GL and 4GL development tools. Then in the early 1990s, companies began purchasing preintegrated sets of OLTP applications to support common back- office functions, such as accounting, manufacturing and human resources. These enterprise resource planning (ERP) applications mushroomed in popularity as companies raced to reintegrate their enterprises and replace old, proprietary or non-Y2K-compliant legacy systems.
Companies are now repeating this build-to-buy evolution on the analytic side of the business. In the early 1990s, many companies developed analytic applications from scratch by deploying a data warehouse and building reports using query, reporting and OLAP tools. Today, vendors are beginning to sell packaged analytic applications that can be deployed in weeks rather than months required for custom-developed solutions.
In many cases, these new packaged analytic applications are mirror images of the ERP packages companies have deployed to automate and integrate back-office and front-office operations (see Figure 1). In essence, the new analytic applications shed insight into the effectiveness of their core operations, enabling them to spot opportunities and problem areas and make quick adjustments. Many companies are discovering that analytic applications are critical to succeeding in today's global, information-intensive economy.
Figure 1: Today's analytic applications are effective in front- and back-office environments.
Just as the ERP market spawned large and powerful software firms such as SAP and PeopleSoft, many vendors are crafting suites of cross-functional, packaged analytic applications in an attempt to stake out a leadership position in this emerging market. Early leaders include Hyperion Solutions, IBM, NCR, Oracle, SAS Institute and Sybase. There are also a host of vendors targeting individual functional areas, such as supply chain, human resources, balanced scorecards and revenue optimization.
Market Size and Growth
According to DataQuest, the market for packaged analytic applications will grow from $602.3 million in 1998 to $3.2 billion in 2003. The fastest growing and largest segment of this market will be CRM-based analytic applications which are expected to generate $724.6 million in sales by 2002, growing at a 58 percent compound annual growth rate. Vendors, such as Broadbase Software, E.piphany, MicroStrategy and Sybase have already staked out leading positions in this hot market segment, with many others soon to follow.
Most of these vendors have also deployed e- business analytic applications or e-intelligence applications to help companies track and understand customer behavior across both electronic and traditional channels. The Patricia Seybold Group expects the market for e- intelligence to grow as fast as the market for CRM analytic applications.
Vendors are currently taking two approaches to building packaged analytic applications. Some are building these applications from scratch by hiring domain experts and conducting substantial research into user and market requirements. Comshare (finance) and VIT (supply chain) are examples of this approach. Most others are acquiring existing applications and retrofitting their infrastructure (if needed) underneath.
What is a Packaged Analytic Application?
Packaged analytic applications are as varied as the vendors selling products in this space. Some vendors define a packaged analytic application as a set of canned reports running off of a predefined data model. Although this configuration may accelerate time to market to some degree, it is more of a package than an application.
The Patricia Seybold Group defines a packaged analytic application as a software-enabled process that walks end users through the life- cycle steps of analyzing data, collaborating with colleagues and taking action. The closed-loop environment empowers users with the information needed to evaluate and analyze business events and opportunities, tools to model and exchange ideas with colleagues, and the authority needed to make decisions and act on them (see Figure 2).
Figure 2: The Business Intelligence Life Cycle
Underlying a packaged analytic application are numerous components (see Figure 3). These include a predefined data model, data acquisition tools (extraction, transformation, load), administrative and meta data utilities, and a business intelligence tool with canned reports, predefined database views and a process-oriented graphical user interface.
Components of a Packaged Analytic Application
Figure 3: The components depicted in blue represent those most typically found in the current generation of packaged analytical applications.
To deploy the application, companies need to tweak the data model, metrics and database views to support their users' unique analytic requirements. They then need to build the necessary extraction feeds to move data from one or more operational systems into the data mart. From there the application's administrative and meta data utilities should make it easy to monitor user activity, automate jobs and synchronize updates.
Since all these components are preintegrated, companies can deploy packaged analytic applications in a fraction of the time compared to building the application from scratch (see Figure 4). Depending on the number of data sources used and the amount of customization required, a packaged analytic application can reduce deployment times by 40 to 60 percent. Moreover, the vendor assumes responsibility for upgrading the underlying technology and application.
Time to Deploy
Figure 4: Packaged analytic applications reduce the time required to deploy an application.
Some packaged analytic applications are designed for use with specific source systems, such as SAP R/3 or PeopleSoft. Source-specific packaged analytic applications can be deployed even faster because they predefine extraction, transformation and load (ETL) programs. For example, Acta Technology's RapidMart for SAP contains predefined ETL templates for populating a sales data mart with data from SAP R/3's sales and distribution module. These types of analytic applications designed to supplement SAP's lackluster reporting capabilities can reduce implementation by as much as 70 to 80 percent.
On the downside, packaged applications may not have all the requisite functionality users require, forcing companies to customize the package. Extensive customization can create problems when the customer seeks to upgrade to the next version of the application. In addition, packages may also bundle specific products databases, modeling tools and business intelligence tools that don't match a company's internal technology standards or architecture. To deploy these tools, companies may need not only to bend the rules, but also hire or retrain staff in these new technologies.
Ideally, packaged analytic applications embody the best practices of top performers in the marketplace. The best packaged analytic applications contain metrics and contextual information that top performers use to analyze events and trends before taking action. For example, Brio uses this approach when helping customers deploy its revenue optimization application called Brio.Impact.
A vendor needs to have considerable domain expertise to build this level of knowledge into an application. In essence, a well-designed packaged analytic application is an expert system that guides novices through a well-formed process for taking intelligent action based on available information. Companies should be able to enhance these expert applications with the best practices of their own top performers.
In addition, a packaged analytic application should contain "hooks" into other programs to facilitate collaborative decision making. Since good decisions are rarely made in isolation, packaged analytic applications need to let users exchange information, record discussions, route and store relevant documents, and perhaps even poll colleagues for their opinions. Many new intranet management products (i.e., corporate portals) excel at fostering this type of collaboration.
Finally, the best analytic applications close the loop, allowing users to make decisions based on the results of their analyses. In many cases, closed-loop processing involves inserting a hook back into an operational application. For example, a marketing manager can use a packaged analytic application to segment customers and score them based on various attributes (profitability, lifetime value, propensity, etc.) They can then use these segments and scores within a campaign management or e- personalization tool to drive effective marketing offers via direct mail, telemarketing operations, e-mail or the Web.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access