Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

November 15, 2002

Outward Bound

Why proactive business intelligence is a hallmark of the real-time enterprise

by Justin Langseth & Nithi Vivatrat

Continued from Page 1

In the same vein, an alert to a store manager about low inventory levels for a particular item might facilitate the approval of an order to be sent to the supplier. There's also the opportunity to facilitate collaboration between alert recipients. In our example, it may be useful for the regional manager to talk with the category manager to jointly plan a strategy to resolve the problem.

The purpose of these alerts could also be merely to ensure that the appropriate personnel are notified about an important condition. In this case, the required action in the workflow could be the simple acknowledgment of the alert. Other actions could include forwarding the information to someone else, flagging as unimportant, or increasing the threshold before which future alerts are triggered. In choosing these follow-up options, users can help the system learn over time, as we will see in the following section.

Workflow options and rules can also be enabled in this type of environment. For example, if an alert isn't acted upon within a certain timeframe, it may automatically escalate to a superior. Or, if recipients reply that they will take action on an alert, they may need to follow up with more detail on the status of the resolution within a certain timeline. By borrowing best practices from customer support workflow systems, managers can quantitatively measure the resolution of issues generated by the BI system and monitor improvements over following periods.

Automatic Learning and Refinement

Having the BI system automatically detect anomalous conditions within a data warehouse reduces the initial setup work; however, the system requires a mechanism to measure effectiveness and learn from experience. In the previous section, we described how the recipient might want an alert forwarded to someone else. The next time the same or similar condition occurs, the BI system will know to send the alert to the other person. If the recipient changes the alert thresholds or flags the alert as unimportant, this provides the system with information to adjust its profile of anomalies and exceptions.

What Is Proactive BI's Role in BAM?

The term "business activity monitoring" (BAM), which was recently introduced by Gartner Group, in general relates to the next evolution of business intelligence. BAM implies the tighter integration of real-time BI and operational systems to let the operational systems become smarter and executives have real-time visibility into core metrics.

Proactive BI represents an important component of BAM, and in many ways goes beyond it. By combining data mining with proactive alerting and collaborative problem resolution, businesses can become knowledgeable, truly think collectively, and react to market and competitive forces in real time.

Although both proactive BI and BAM as a whole are in the early stages, once one company adopts the techniques to gain a competitive advantage, other companies in its market are sure to quickly follow.

This type of continual feedback, learning, and improvement allows the system to tune itself over time, so that after a few weeks the best possible information is provided to the proper individuals, and the number of false alarms is as low as possible. The proactive BI system can become a constant monitor of the business at all levels.

Eventually, users throughout the organization will begin to trust the system, and the absence of an alert will be taken as a reassuring sign that everything is operating within normal parameters. Users who once spent time running daily reports can spend their time resolving problems instead of manually gathering information. Over time, the proactive BI system will become a critical part of the organization's day-to-day existence, a true BAM system (see the sidebar, "What Is Proactive BI's Role in BAM?" at right) and a true "best friend" to all levels of the organization.

But Why Wireless?

A large part of being proactive means reaching out and communicating proactively with people across the enterprise or supply chain who need to either be aware of, or take some action relating to, a business condition identified by the data warehouse and BI infrastructure. Although email is a ubiquitous and obvious way to distribute information, information and alerts that are truly business-critical and time-sensitive are best communicated wirelessly — for example, directly to a user's mobile phone, pager, or wireless PDA in the form of short text messages. (The last thing you want is for a critical business alert to be lost in a sea of spam or other less critical email traffic.) Wireless also offers the benefit of delivery status, guaranteed delivery, and the ability to take action immediately through two-way SMS, WAP, J2ME, or other technologies.

Making it Work

The best way to move from a passive and reactive data warehouses to proactive BI (perhaps with a wireless component) is to work in phases. The first step may be to enable simple alerts to a small group of individuals, based on simple measures such as budget variance, sales plans, or inventory levels. For example, a store manager may want to be notified on a wireless device when inventory levels of any popular product are exceptional high or low.

When the concept is validated with a test group and true value is demonstrated, the principles of proactive, wireless BI can be extended to larger and larger audiences. In terms of off-the-shelf products to address the needs of proactive BI environments, no end-to-end solution is available yet. However, separate components, such as wireless rule-based alerting systems, anomaly detection data mining tools, and real-time extract, transform, and load, and enterprise application integration tools do exist and can be integrated to enable proactive BI.

Rate This Article

Comments:

Optional e-mail address:

Over the next few years, we can also expect a new class of BI tools to emerge that are built around the core tenets of proactive BI. These new tools, combined with the latest wireless devices, will allow for some very impressive (and useful) systems to be implemented. Smart enterprises will use automatically generated real-time alerts combined with follow-through capabilities to accelerate workflows and dramatically increase the value of existing BI systems. By leveraging BI systems in this way, real-time decision support will become a core component of almost every employee's daily activities, causing efficiency gains and significant benefits to the bottom line.

Justin Langseth [langseth@claraview.com] is a founder and CTO of Claraview LLC. Claraview provides system architecture, project management, and implementation consulting services to companies and government agencies implementing real-time, proactive, and wireless-enabled data warehouses and alerting systems. Prior to Claraview, Langseth was the founder and CTO of Strategy.com

Nithi Vivatrat [vivatrat@claraview.com] is a founder and managing director of Claraview LLC. Prior to co-founding Claraview, Vivatrat was the director of product management at Strategy.com

RESOURCES

Gartner Research Note COM-14-9785 (March 18, 2002) contains Gartner's view of business activity monitoring's business value. See www3.gartner.com/1_researchanalysis/focus_areas/aim/aim040302/aim04032.jsp for more details.

Related Article at IntelligentEnterprise.com: "Agility Training," April 28, 2002