When you think of business intelligence (BI), what do you think of? From experience, many of us think of massive databases filled with high-quality, integrated data just waiting to be twisted and turned in a myriad of analytics. We appreciated this pristine world and its separation from the chaos of operations.

Historically, operations were completely separated from the analytic BI environment for good reason. Strategic queries, involving hundreds of thousands of records, would have destroyed the performance of critical operational systems, responsible for single record transaction processing. Until recently, technology was not available to support a “mixed work load” (i.e., the ability to accomplish massive analytical queries and short, transaction processing at the same time).

Certainly, that was the focus of business intelligence in the past – but no longer. Business intelligence has “invaded” the operational space in a big way, offering in-line analytics, real-time or near real-time decision-making support for all employees in the enterprise. Today’s BI environment includes three forms of BI – strategic, tactical, and operational.

As enterprises saw success in their BI environments, they began demanding that the same features – integrated data, easy-to-use access tools, affordable storage and database capabilities, etc. – be made available for the operations of the business. This was a major shift from traditional business intelligence, and another hurdle for most technologists and BI vendors. It meant that the traditional BI environment had to expand to handle all forms of BI with acceptable performance for all users – analysts, executives, line of business managers and, now, even front-line employees.

Of necessity, the BI implementers studied and evaluated the differences in the three forms of business intelligence to ensure an appropriate BI architecture and technological environment. Table 1 illustrates the characteristics and differences between strategic, tactical and operational BI. The business focus, users involved, currency of the data, and even the types of data used differ significantly, depending on the type of business intelligence.

Table 1: Characteristics of Strategic, Tactical, and Operational BI

So – Where’s the Pain?

Certainly, operational business intelligence requires real or near-real time data for some of its queries and analysis. But the more we speed up the data acquisition and integration process to support operational goals, the more complex the environment becomes. Let’s face it, putting up-to-the-second data into the hands of the entire corporation is expensive and creates a major burden on IT.

We need to address the real-time scenario with a more comprehensive and pragmatic approach. Smart corporations require a detailed cost-benefit analysis from their business users and IT to determine just how fast is fast enough. Perhaps what corporations need is not universal real-time data delivery, but rather “right-time” or “on-demand” data delivery that is a mix of real-time, near real-time and historical data.

On-demand data delivery consists of a mix of instantaneous, rapid intermittent, and longer batch-type processes – each yielding a different timeframe for its data: sub-second to a few seconds to several hours to overnight or longer. For example, a securities trader will need immediate access to stock market data. The credit card approval process may appear to have immediate access to customer data; but, in fact, the process is complex and requires a few seconds for approval to occur. Order fulfillment information may be generated once or twice a day, and mailing lists may be generated once a month or less often depending on the timing of marketing campaigns.

These are right-time deliveries of data that are completely appropriate for their particular processes. Yet, to the employees using them, they appear to be real-time. In reality, the applications generating these bits of data most likely use a mixture of real-time, sporadic and longer-cycle data delivery processes.

Acceptance is the First Step

If you accept the premise that right-time data delivery is the appropriate direction for your enterprise, the challenge then becomes to properly identify the ideal time continuum for each business process – that is, which processes need to be accelerated and why.

Here are a few examples of operational processes that desperately need a shot of business intelligence:

• You may need to speed up data delivery if your enterprise is at a competitive disadvantage because its decision-making processes are too slow. Being able to quickly respond to a competitor’s new marketing message, counter negative product information or retain a customer who is about to leave are operational processes that require a BI injection.
  
  
• The time it takes from order to fulfillment may be off the mark, and the enterprise loses business to companies that can deliver faster. The ability to tie customer demand to a company’s manufacturing schedule requires a BI injection.
  
  
• Reducing the latencies between when a significant business event occurs (such as a campaign that results in an inventory outage) and when the enterprise takes action (such as ending the campaign or changing the product mix for it) may be “just what the doctor ordered” for your operational BI project.

The Diagnosis

The next step is to perform an honest baseline assessment of your existing data delivery capabilities (e.g., available technologies, maturity of the BI architecture, existing personnel, etc.), and to combine it with a solid understanding of the business requirements for right-time data. The assessment will uncover any weaknesses in your technology, architecture, resources, processes, etc. It is critical to understand that these weaknesses will only be exaggerated as you speed up the enterprise.

The technology assessment consists of documenting what is in place, what can be sped up, and what new technology will be needed to increase data velocity. Data integration tools (e.g., EAI, EII and ETL), database characteristics, data quality software, networks, and even query and data access tools should be included in this assessment.

The determination of the business requirement may be more difficult. Start with a solid, detailed definition of what right-time data delivery is to ensure business community understanding. Once the definition is socialized and accepted, you can then develop scenarios that demonstrate how operational business intelligence can be combined with operational processing to create a smarter enterprise. Look for gaps in current operational processing where operational BI could help, and develop a detailed scenario demonstrating how the new process will help.

Taking your Meds

BI implementers can use these types of case studies to establish which components of their BI architecture should be altered, replaced, or rebuilt, and ensure faster delivery and integration of BI capabilities into operational processes.

Delivering analytics to operational personnel can occur in a number of ways. For example, you may be able to embed your BI applications or their components into the actual operational application. In this way, you can facilitate in-line analytics within the operational system itself. If that is not feasible, an alternative approach may be to call on a BI service at the decision point. Service-oriented architectures (SOAs) are required for this approach. In either case, you must create an operational BI application by combining the data-centric capabilities of business intelligence with the process-centric capabilities of the operational environment.

Both approaches may deliver only an outward appearance of operational BI. The real challenge will be to build out a solid data integration framework or infrastructure that fully supports the front-end delivery of business intelligence and operational information to the operational personnel. It is important to overcome the tendency to solve a very specific problem and, rather, create an enterprise IT strategy for operational business intelligence.

Recent articles by Claudia Imhoff

• Big Data in the Operational World Requires a Different Mind-Set: A Spotlight Q&A with James Taylor
• Bridging IT-Business Divide: A Two-Way Street
• Collaborative BI Sets Stage for Collective User Action
• Split Personalities Put Data Scientists in Play