Agile Organizations Correlate Business Events

A typical organization generates millions of events every day: systems, applications, employees, customers, partners, or any stakeholder that interacts with the organization usually generates these events.

In isolation very few of these events are of interest; however, when analyzed and correlated, these events provide information rich in business context and ideal for decision-making.

An organization’s agility is directly tied to its ability to identify important business events and notify the right decision-makers within the window of opportunity where action makes all the difference.

As organizations move toward a real-time enterprise, a growing wave of technology aims to provide the complex event processing, analysis and display/notification for IT to deliver this perfect real-time decision support system.

Have you considered how your organization’s customer service, sales, logistics or accounting departments might have acted differently had they possessed better-quality knowledge when they needed it?

Unfortunately for most organizations, the recognition of important business events happens after the fact, if it ever does. The challenge is that, as you get closer in time and space to where an event is taking place, the bigger the context and significance of that interaction.

Some examples of this include:

  • The total cost of sales resulting from the successful up-selling of a customer to another product may be very different depending on whether the original order has already shipped.
  • A broker making trades on her personal account might have very different significance to insider trading investigations depending on whether one of his/her clients execute a similar or related trade shortly before or after that transaction.
  • A refrigerated truck on its way to Store A carries stock that is in heavy demand at Store B, which is close to this route.
  • To overcome the latency issues of learning about such events after the fact, an organization needs to improve its environment to start taking advantage of business events.

    A simple methodology to get started includes:

    1. Capture useful events and then turn them into useful decision-making aids, whether individual events or rolled-up events.

    2. Route the events to the right person, application, or business process that could make use of them.

    3. Process events on their own and in conjunction with other information.

    4. Keep interested parties aware in real time and when they need to take action.

    Capturing Business Events

    Fortunately, it doesn’t take a major redevelopment of your applications infrastructure to enable your organization to capture business events and act on them. It’s actually very simple and can be done using an existing database and enterprise application integration (EAI) or message-oriented middleware (MOM)-based technology.

    Events can be raised at many levels of the organization — systems monitoring, enterprise resource planning (ERP), customer-relationship management (CRM), point of sale, and so on.

    The events carry much of the information required to correlate them: where they originated, when they happened, and what happened before and after them. Plus, you need to track the event data itself (for example, the purchase order details when a new purchase order event is captured).

    Business Event Processing

    After recognition, the next important step is determining what action to take. If you really have the ability to propagate an event or series of events, then it must be important for something to be done about it. One example of a composite event is a business-to-business (B2B) order transaction.

    Most B2B order transactions involve many events (order request sent, order request acknowledged, and so on). It’s the combination of these events that most businesses care about.

    There are two paths to identifying when something relating to your B2B orders needs action — event metrics and key performance indicators (KPIs). The first is measuring events at the instance level.

    You may want to know whenever a particular order is about to break a service-level agreement (SLA) because of a vendor’s inability to supply a part. The second is based on a KPI on aggregated event data. You may want to know when a vendor’s average response time to an order request becomes greater than two hours.

    Automated Response and Notification

    After capturing the business events and processing them, your organization is capable of identifying the right situations in which to take action. The trigger events — usually when SLAs are about to be broken or when suspicious activity is identified — normally require human intervention or notification.

    In more of an escalation situation, it is likely that many different people, services, and supporting processes might be involved. By directly engaging all the affected parties immediately, the organization ensures that the resolution occurs more smoothly, has a higher success rate and is more cost-effective.

    Being able to leverage organizational knowledge is a major source of competitive advantage and an essential ability of an agile organization.

    Event correlations can better identify the “real” or the root (as opposed to “apparent”) cause of an outcome and thus can improve enterprise business activity monitoring (BAM) applications or help in tactical “predict and improve”-type tactics.

    Harnessing the power of event correlation and analysis provides for a much more deterministic approach to seeing into the future and driving a predictive real-time enterprise.

    John Deeb is a director in the product management team for Oracle Fusion Middleware, a complete, integrated middleware platform. His areas of focus include enterprise integration, BPM and BAM.