Continue in 2 seconds

What is Active Business Intelligence? Part 2

  • May 01 2002, 1:00am EDT

Last month, we introduced the concept of active business intelligence (ABI) as having the characteristics of:

  • Supporting tactical decision making,
  • Leveraging actionable intelligence, and
  • Enabling the learning cycle.

This month's column continues by describing the second and third characteristics of ABI.

Leveraging Actionable Intelligence

The second characteristic of ABI is that it enables complete BI, which bridges the gap from data to action.1 Complete BI consists of the following five stages:

  • Observe – What is happening?
  • Understand – Why has it happened?
  • Predict – What will happen?
  • React – What should we do now?
  • Reorganize – How can we do it better?

The observe function captures the history of the business and is the bread and butter of data warehousing. The understand function comprehends the dynamics of the business, as inferred from the data. The predict function forecasts the future state of business variables based on a model of business dynamics. The react function decides and executes a course of action based on an understanding of the business dynamics and predictions about future trends. The reorganize function learns to improve business processes by refining best practices. Improving a single decision is not sufficient; improving the decision process is the ultimate objective.

Traditional BI often stops after the first three stages. Its responsibility unfortunately ends at the pixels on the screen or the beeps on the pager, failing to insure real benefits for the business.

In contrast, ABI moves us beyond those initial stages into the fourth and fifth stages. ABI enables us to react to the current business demands, threats and opportunities based on our observations, understanding and predictions. Also, ABI enables us to reorganize to meet future demands, threats and opportunities.

ABI employs the following mechanisms to affect action:

  • Actionable reporting – periodic reporting with highlighted exceptions,
  • Actionable analysis – ad hoc analysis with suggested actions,
  • Actionable alerting – alerting directed to the right person at the right time, and
  • Activated processes – embedded analytics.

The first mechanism is traditional periodic reporting with a twist. Actionable reporting implies a summary of observations that highlights exceptions (i.e., situations that are abnormal) and suggests actions (i.e., ways of correcting those abnormalities). Hence, actionable reporting involves more than the first stage of observing; it requires the understanding of business dynamics and the prediction of future events. Instead of the reams of green-shaded paper, actionable reporting is the substance of a personalized workplace within an enterprise portal.
The second item is traditional ad hoc analysis, but with a twist. Actionable analysis implies a synthesis of diverse facts using complex algorithms, just in time, as business events require. Again, exceptions to the norm are highlighted, and corrective actions are suggested.

The third item is alerting, but in a universal context. Actionable alerting implies the continuous monitoring of business events, predicting situations that will become adverse to the business and suggesting actions to avoid the undesirable. In addition, actionable alerting must identify the individuals who should perform the actions and choose the proper way to communicate the alert (e.g., PDA, beeper, etc.). With an intimate knowledge of the organizational structure, actionable alerting requires a universal context for alerting anyone, anywhere, anytime.

The fourth mechanism is embedding analytics directly into business processes, possibly with little human intervention. Activated processes mean incorporating the full intelligence of ABI into the minute-by-minute activities of the enterprise. In other words, ABI activates business processes by fully leveraging business intelligence. For example, activated processes are Web sites that recommend products for which the past behavior of a customer indicates a high probability of acceptance.

There are three requirements for leveraging actionable intelligence.

First, the greatest potential for actionable intelligence is to embed well-designed analytics into key business processes. The principle is to reduce the distance from insightful intelligence to effective action. At the same time, do not fragment the enterprise architecture with piecemeal analytics that do not fit into a coherent framework.

Second, successful ABI implementations require just-in-time latency, not necessarily zero latency. One should invest in the infrastructure of data connectivity and integration to the extent warranted by actionable intelligence. The analysis should be reported when action must be (or can be) taken, which may be some duration after a transaction commits. This is a relative time scale dependent on the ABI learning cycle (explained momentarily). As your business culture learns to absorb ABI capabilities, the latency to satisfy the just-in-time requirement should become less and may approach zero. An architecture that reduces and even eliminates extract, transform and load processing is then required, thus integrating operational and informational data into a "single-version-of- truth" environment.

Third, ABI is not the same as closed-loop data warehousing, implying that analysis results must feed back into operational legacy systems. For some ABI implementations, this may be the case. In general, analysis results must be directed to specific individuals having responsibilities for specific processes. Legacy systems may be on that path. Or, new paths may utilize wireless technology and enterprise portals, bypassing legacy systems entirely.

Enabling the Learning Cycle

The third characteristic of ABI is that it enables a learning cycle that can incrementally improve business benefits. Because ABI impacts people and their job responsibilities, changes will be slow and incremental.

In Figure 1, the learning cycle of ABI starts with the ABI technical capabilities, which produce greater insights into the business. This positively impacts the business, which then motivates the business to increase the ABI capabilities in specific ways to further impact the business, and so on.

Figure 1: The Learning Cycle of ABI

There are two conclusions that become obvious when enabling the learning cycle of ABI.

First, successful ABI implementations are highly iterative. Start simple, and let the organization slowly absorb the greater insights from the ABI, striving to realize tangible business benefits.

Additionally, successful ABI implementations are never-ending. Keep the cycle going! As tangible business benefits are realized, learn from the experience by celebrating the successes and correcting the failures. Invest in additional ABI capabilities that will yield specific insights into the business dynamics. Then realize specific benefits from those insights.

In summary, consider your BI systems today. To what extent do they support the react and reorganize functions? How would your business benefit if these functions were supported better? What are several tangible steps that move your BI systems toward supporting more ABI?


1. More details on complete BI are available in: R. Hackathorn, DM Review, November 2001.

Author's note: Support for this work came from Teradata, a division of NCR, as part of their investigation into the value propositions behind Active Data Warehousing.

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

Don't have an account? Register for Free Unlimited Access