How often have you thought, "If I only knew then what I know now"? Have you wondered how your sales, logistics, or customer service personnel might have acted differently with a customer or supplier had they been armed with better knowledge? As the figure highlights, people tend to recognize important business events, scenarios, and information after the fact, if ever. Often, this information is tabulated long after the event has occurred in the form of customer churn, lost deal analysis, customer dissatisfaction, or reports. Consumers of those reports are often only senior managers and executives who then exhort their teams to do better next time.

The problem is that, as you get closer in time to when an interaction occurs, its larger context and significance aren’t immediately apparent, or it’s significant only in relation to other events happening in the organization. For example:

- The total cost of sales resulting from successful up-selling of a customer to another product may differ, depending on whether the original order has already shipped. You want the increased revenue, but the particular offer you make might change, based on this fact.

- A broker making trades on his personal account might have different significance if one of his clients executed a similar or related trade shortly before or after his transaction.

Business Activity Monitoring (BAM) offers an approach that can highlight an important event or scenario as it’s happening and relay that information to the employee. Information can be distributed through simple alerts, input into existing applications, or delivered as updates through a dashboard. This lets the employee in the midst of the event become aware of the bigger picture and respond appropriately based on its significance.

"Doing the right thing when it makes a difference" is often what separates the best companies from their competitors. To do so, you need two key capabilities. You must:

- Know there’s a decision to be made

- Be able to make and execute that decision.

The first capability involves identifying that an important event or scenario is happening. This may not always be evident to an employee without additional information. The second capability involves giving the employee the tools and information to do something about it (or having an automated process occur, if that's more appropriate).

BAM provides the infrastructure that can help organizations put these two capabilities in place. BAM is often characterized as only a real-time dashboard, but this technology space is rapidly evolving to becoming a critical element of a larger technology foundation that incorporates Business Process Management (BPM), business analytics, Service-Oriented Architectures (SOAs), and Event-Driven Architectures (EDAs). BAM provides the critical junction at which events, services, and processes are linked with rules, notifications, and people.

Responding Effectively
The next step is determining what action to take. If you can propagate an event (or series of events) from the source to a person or process, they should be able to act on it. For this to occur, the person or process must have the necessary information to make a decision, the authority to enact it, and the tools and resources needed. Truly providing this empowering environment requires a framework in which applications, rich services, and structured workflow can be easily incorporated. Simple distribution of alerts and notifications is one type of response, but more is needed.

Among the challenges in enabling someone to respond effectively is clarifying what possible options exist. Because the triggering event probably isn't a regular occurrence, the employee may not know what needs to be done without supporting information and tools. Often, it might be better for an employee to have a structured workflow that frames the options, and also automatically moves the event, and resulting decision, through a complete process to conclusion. If this is a call center escalation situation, many different people, services, and supporting processes might be involved. By immediately directly engaging all the affected parties, the action or resolution will typically occur faster and more smoothly, cost effectively and successfully. In contrast, trying to do all this later would probably involve additional time, resources, and people, and be less effective overall.

So a platform that supports all these capabilities must do more than just provide monitoring, visualization, and notification, although those capabilities remain important. Fortunately, infrastructure and approaches contained within service-oriented computing and event-driven computing help deliver this platform.

BAM as Bridging Technology
As we've been discussing events, services, processes, workflows, notifications, and people, we've introduced concepts that underpin service-oriented computing, Web services, and event-driven computing as enablers of this overall solution.

- Service-oriented computing is defined by a request-response paradigm. Individual services are invoked, or requested, to do some work. In this model, the sender pushes the request to the service and the target service is well-known by the sender.

- Event-driven computing is defined by a producer-consumer paradigm, where these two are decoupled and a producer doesn’t explicitly know who the consumer is. In this model, services consume, or pull, events that trigger some action.

These two computing models complement each other and together provide a flexible, efficient approach. In cases where it's well-known what particular actions need to occur, the explicit invoking of specific services or processes should be used.

BAM is an important bridging technology since it incorporates event capture and processing capabilities, combines them with business analytics and logic, and then invokes visualization, notification, or services/processes in response to particular scenarios. BAM provides both the recognition and response capabilities necessary to better empower operational resources. It has one foot in event-driven computing, with its ability to capture and process events, and the other in service-oriented computing, via its innate ability to seamlessly monitor services and processes and then invoke them as necessary.

BAM Grows Up
BAM grew out of the application integration and messaging technology space as a way of tapping into these transaction flows to provide visibility and monitoring, and complement existing Business Intelligence (BI) and analytics solutions. However, BAM has increasingly provided unique and more general capabilities through its ability to capture and process not just transactions, but also any events as they occur. An example of this is the increased adoption of technologies such as Radio Frequency Identification (RFID), which can provide real-time event feeds outside application transactions. Similarly, BAM technology continues to provide a bridge between back-end processing and higher-order business metrics and rules.

Some key capabilities of a BAM solution are:

- Event collection: Seamlessly consumes events produced by EAI, BPM, Web Services, and orchestrated processes; can also consume events from messaging, database, system, sensor, and other sources.

- Event processing: Derives significant events from the raw event stream such as correlation, aggregation and filtering.

- Analytics: Processes the data and events in real-time, leveraging existing BI data and tools to provide interpreted information.

- Access: Distributes the information to all appropriate people, applications, and business processes, anytime and anywhere, that have registered interest.

- Response initiation: Automatically triggers the required action to events and information automatically; allows for business optimization and management by exception.

One key BAM capability we’ll discuss further is event processing and the ability to derive events from more basic, lower-level events. This area is often described as Complex Event Processing (CEP), but to fully explore that topic would require a separate discussion. In this article, let's focus more on how to use that capability to derive events in the context of BAM and integration.

Why Focus on Business Events?
Focusing on events, not just data, is important for distinguishing this range of solutions from typical reporting, analysis, and other after-the-fact information processing. While supporting data is often necessary for putting an event in context, a key difference between an event and data is that an event indicates a change of some sort. This change indicates something has just happened and this immediacy of knowing that change of state is an important characteristic of an event. Visibility to that event can identify a triggering business situation that will initiate other actions and pull the necessary information along with that event.

An important distinction must be drawn between events in which we're ultimately interested, and the raw system and application events are generated and collectively indicate an event is occurring. Higher-level events are used to trigger business responses or notify people. Filtering, correlating, or aggregating lower-level raw events identifies these events.

However, the challenge in efficiently processing events and correlating them is that seeing one event in an identified pattern doesn’t mean you'll see the rest of the events in the pattern. In actual situations, you may rarely complete the correlation. So trying to write code in existing services and processes to identify these complex event scenarios leads to complicated pre-processing logic. It also leads to the inefficient use of system resources as new threads are created and the service or process waits to see if the complex event pattern is completed or times out.

So it makes sense to break out event processing as a specialized task. This allows all the when and what-if logic to be explicitly modeled and managed and deployed onto run-time platforms specialized for that task. Following this approach, we can again have services and processes that are focused on doing work, rather than trying to figure out if work should be done. In this scenario, dedicated event processors consume raw events and emit well-defined, higher-level events, which can be distributed to people, services, and processes to be acted upon.

Linking It All Together
To get the greatest return from our technology investments, it's necessary to create the vital links between events, information processing, and processes. Increasing the focus and adoption of architectures, such as service-oriented and event-driven computing, will make it easier to achieve these links. To accomplish this:

- Identify critical business interactions and determine how they can be leveraged to drive increased revenue or productivity if better information were available.

- Expand your thinking about information delivery beyond simple alerting and monitoring applications to include automated processes and workflow applications, so there's minimal separation between issue identification and resolution.

- Leverage modern application platforms that let you easily combine events, processes and services to deliver integrated solutions.

Taking a broader perspective of BAM and its vital role in enabling the whole organization will let us take a big step toward using information to truly run our businesses more efficiently.