Portals eNewsletters Web Seminars dataWarehouse.com DM Review Magazine
DM Review | Covering Business Intelligence, Integration & Analytics
   Covering Business Intelligence, Integration & Analytics Advanced Search

View all Portals

Scheduled Events
Archived Events

White Paper Library
Research Papers

View Job Listings
Post a job


DM Review Home
Current Magazine Issue
Magazine Archives
Online Columnists
Ask the Experts
Industry News
Search DM Review

Buyer's Guide
Industry Events Calendar
Monthly Product Guides
Software Demo Lab
Vendor Listings

About Us
Press Releases
Advertising/Media Kit
Magazine Subscriptions
Editorial Calendar
Contact Us
Customer Service

Delivering Intelligent Operational Applications (or How to Move Up the Value Chain)

  Article published in DM Direct Special Report
August 16, 2005 Issue
  By Munish Gandhi

Current business applications fall into one of the following classes:

  • Operational Applications: These applications form the operational backbone of the company and are responsible for functions such as billing, order management, order provisioning or call center management. These systems store and manage information in a system of record and need information accurate to the last second of activity on a particular customer. In other words, these systems have low latency but are low in intelligence.
  • BI Applications: These applications form the analytical backbone of the company and serve a strategic need to understand the company's customers, products and services. These systems store and manage information in an analytical data store such as a data warehouse, but are designed to rely on information that is at least a day old. In other words, these systems are high in intelligence but have a high latency.

A third class of applications is now emerging. We call these apps - intelligent operational applications, or just IO apps, as shown in Figure 1.

Figure 1

  • IO apps are integrated into the day-to-day operations of a company.
  • IO apps are smart applications that increase the availability of intelligent information for hundreds (if not more) of small operational decisions that improve the customer experience or improve profitability.
  • IO apps are designed for use by a large number of front-line personnel or in transactional interfaces such as the Web for immediate action. (Contrast this with BI applications that are designed for use by a few business analysts.)
  • IO apps merge information from analytical stores with up to-the-second information from the operational stores.

Traditionally, it has been difficult to measure the value of a BI application because the effect of such applications is indirect. An IO app, on the other hand, has a direct impact on revenue and customers. This allows the ROI of an IO app to be measured and proven easily. Further, as the organization becomes dependent on these applications for basic operations, intelligence applications move up the value chain to establish themselves as critical applications for running the business.

What is an example of an IO app? What are the essential characteristics of a useable and useful IO app? What capabilities distinguish an IO app from other apps? What is the best architecture for an IO app? We have been able to answer some of these questions based on our experience implementing several IO apps. We share our insights in this article.

Example: Transforming Up-Sell through an IO Application

IO apps may be thought of as adding immediacy to traditional BI applications. Alternatively, they may be thought of as adding intelligence to traditional operational applications. However, IO apps are best thought of as enabling a dramatic transformation of existing business processes.

Consider up-sells. Imagine a customer who has just purchased a camera. At this point, what is the best up-sell for the customer? Should it be an extended warranty, extra batteries, extra lenses, photo printer, software for managing a digital library or 20 other items that are offered when we make purchases on a Web site?

Operational up-sell applications, such as up-sell offers on a Web site or thanks-for-your-order e-mails, usually throw up all possible up-sell information regardless of the relevance to the customer. A natural reaction of the customer is to ignore the full package of up-sells.

On the other hand, a BI-powered up-sell application such as an e-mail campaign is usually too infrequent to take advantage of "hot" opportunities. These BI-powered campaigns are not event-based and thus miss the relevancy of the opportunity.

An IO up-sell app can transform the up-sell process:

  • An IO up-sell app can assemble both analytical and event information in real time to create the best offer for the customer. For instance, knowing that the customer has purchased an extended warranty in the past would increase the propensity of the warranty offer. Alternatively, a purchase of a high-end computer would increase the propensity of a software add-on.
  • Once an up-sell is identified, the same up-sell can be consistently offered across the full range of channels - on the Web, through -mail or a call center.
  • In addition, the same upsell can be consistently offered across time. It can be offered immediately on the Web, an hour later when the customer calls the customer service center or that night through e-mail.

Notice the advantages of an IO app over traditional BI or operational applications:

    1. The offer is more intelligent and, hence, more interesting to the customer than traditional offers.
    2. The offer is immediate and, again, more interesting to the customer than traditional campaign offers.
    3. The offer is consistent across channels.
    4. The offer is consistent across time.

Essential Characteristics

An IO app needs to adhere to the "just-in-time" philosophy. To explain, Dell's supply chain is built around the just-in-time philosophy. Dell waits for a customer to ask for a product rather than build a large inventory of products and then "push" it to the customer. The move to just-in-time thinking:

  • Requires Dell to be agile in responding to a customer's needs,
  • Requires Dell to be smart about processing a customer's request, and
  • Allows Dell to simplify its business model by not having to deal with the complications of a large inventory system.

Figure 2: Dell's example inspires us to demand that an IO app must be agile, smart and simple.


Many of the current reporting systems require that an "inventory" of data in the data warehouse before the data can be used. If the requested information is not in the warehouse, these systems will not be able to answer the business needs. Further, traditional data warehousing development cycles take months to integrate new sources. All this tends to make information "heavy" and prevent it from reaching the business where value can be created.

By contrast, an IO app must ensure that information is "lightweight." That is, the solution must be agile enough to source data from any relevant source:

  • It should be possible to source information from an analytical store such as a data warehouse. For example, the churn propensity of a customer and the customer's purchase history may be sourced from a data warehouse.
  • It should be possible to source information from an operational data store. For example, the status of all the recent orders may be sourced from the fulfillment and billing databases.
  • It should be possible to source information from a real-time source such as a web service. For example, the shipping and delivery status of a customer order may be sourced from FedEx or UPS.


An IO app must be smart and sophisticated enough to drive decisions. The information should be:

  • Streamlined to correlate the information from the different sources,
  • Evaluated to prioritize the important and highlight the relevant, and
  • Probed to discover hidden propensities.

One of the most effective ways to perform this smart analysis in real-time is to drive it using rules that are based on existing best practices in the company (or other companies).


The smart analysis reduces the mass of data into a simple information packet that invites action. The idea of discarding information may be anathema for use by a business analyst, but is essential for an IO app.

In our up-sell example, we gain the trust and confidence of a customer if we present the customer with a few highly relevant offers. If we present the customer with a large list of all possible up-sells, the customer will ignore all the offers. To make matters worse, a business incurs additional costs (time spent by the customer service staff, for example) for making an offer that is not relevant for the customer.

Core Capabilities

For an IO app to be useful, the application must be able to bridge the gap between the information scattered across the ecosystem and the business insights needed by the business to drive decisions.

The capabilities that are essential to bridging the gap may be summarized using the GAP mnemonic (see Figure 3).

Figure 3

  • Gather Data: The information for any decision is scattered across many different internal and external systems in the business ecosystem, the first task of a business visibility system is to gather information from all these sources to create a 360 view of the customer or business process.
  • Analyze Information: Traditional data collection by analysts and business managers is followed by a process of streamlining, scrutinizing, probing and evaluating the data before it becomes usable and actionable. The second major capability of an IO app is its ability to streamline, evaluate and probe the data set.
  • Present Insights: The insights gained from the data gathering and analysis should be delivered to the consumer of this information. The third capability of a IO app is its ability to deliver the relevant and actionable insights via a Web interface for human consumption or via a Web service for consumption by another application.

Architectural Components

Figure 4 illustrates the core architectural components of an IO application.

  • The IO engine components on the left hand side of the graphic provide the GAP capabilities for gathering, analyzing and presenting the information.
  • The IO designer components on the right hand side of the graphic allow the designer components to manipulate the rules executed by the engine components. The independence of the actual rules from the execution of the rules is key to making the solution agile and smart.

Figure 4

The components are:

  • Data Assembler: This component gathers information from various hard sources such as databases or soft sources such as Web services.
  • Assembly Rules Designer: This component specifies the detailed SQL Queries, Web service queries and data set assembly rules that need to be executed to gather the data across the ecosystem.
  • Data Analyzer: This component streamlines, scrutinizes and probes the data so that compact and refined information can be presented to the business.
  • Best Practices Rules Designer: The rules for the data analyzer are derived from the existing best practices in the company. These rules are managed by the best practices rules designer and are fed to the data analyzer for its analysis of raw data.
  • Presentation Web: This component presents the data directly to the user using the web or may deliver the data to another user interface such as a Web site using Web services.
  • Presentation Designer: This component allows for greater control and flexibility in changing the presentation Web.

The move toward "just-in-time" operations is forcing businesses to move towards IO apps. Fortunately, we now have the technology necessary to meet the exacting characteristics and capabilities required of these applications.

The result: an enterprise where business intelligence breaks free from ivory towers of business analysts and strategy developers. This intelligence now:

  • Pervades hundreds and thousands of smart decisions made every day by every employee - all the way from the CEO down to front-line employees who act on behalf of the company to service a customer.
  • Pervades every system from a Web-based front end to e-mails that service a customer.

For more information on related topics visit the following related portals...
Business Intelligence and Database/Application Performance.

Munish Gandhi is a founding principal of Infoveo LLC. Prior to Infoveo, he was a prinicipal engineer and information architect at Network Solutions, CTO of LifeMinders and the director of data warehousing at CNET. He holds three patents and has a Ph.D. in Computer Science with a specialization in data modeling. You can e-mail him at mgandhi@infoveo.com.

Solutions Marketplace
Provided by IndustryBrains

Extract Greater Value from Your Data
Your data is the engine that drives your business. Let the Experts at IBM help you put the information you need right at your fingertips. Click to register for more information.

IBM Master Data Management
IBM Master Data Management provides a single view of critical information by bringing together all core components required for a successful enterprise data management strategy. Learn more!

Analytics for Oracle Applications
Get sophisticated analytics and real-time reporting for Oracle and MFG Pro ERP systems based on a packaged data warehouse. Immediate results from packaged Business Solutions from Jaros Technologies.

Free EII Buyer's Guide
Understand EII - Trends. Tech. Apps. Calculate ROI. Download Now.

File Replication and Web Publishing - RepliWeb
Cross-platform peer-to-peer file replication, content synchronization and one-to-many file distribution solutions enabling content delivery. Replace site server publishing.

Click here to advertise in this space

E-mail This Article E-Mail This Article
Printer Friendly Version Printer-Friendly Version
Related Content Related Content
Request Reprints Request Reprints
Site Map Terms of Use Privacy Policy
SourceMedia (c) 2005 DM Review and SourceMedia, Inc. All rights reserved.
Use, duplication, or sale of this service, or data contained herein, is strictly prohibited.