This article originally appeared on the BeyeNETWORK.
Why does business intelligence (BI) matter? Why does “Operational Business Intelligence (OBI)” matter even more? And why would “Operational Business Intelligence Applications” mean not only the difference between profit and loss, but also between survival and bankruptcy for many organizations?
Business intelligence is neither a product, nor a system. It is an architecture and a collection of integrated operational, as well as decision-support systems and databases that provide the business community with easy access to the organization’s data. Essentially, business intelligence is business success realized through rapid and easy access to actionable information. The information should provide timely and accurate insight into business conditions about customers, products, services, finances, markets and investors.
Traditional Business Intelligence
Traditional business intelligence supports the “Power Users” by providing a set of data access tools for creating fancy cubes, charts, graphs with multiple dimensions with OLAP(Online Analytical Processing). It is often done to track organization’s performance for the upcoming months and years.
In a traditional business intelligence environment, a power user typically tries to predict an organization’s future performance based on the past data. This is done using a set of query, reporting and analysis tools. But what about the organization’s current business intelligence performance? Who is looking at the “real-time” data to make changes in the running of the organization as it happens?
Figure 1: Traditional Business Intelligence
Taking immediate action, which is based on the present information, is not only business intelligence. This is known as operational business intelligence.
- The main goal of operational business intelligence is to offer immediate results by providing information to the operational users. This is done to quickly implement any necessary changes in running of the business processes.
Verifying the ramifications of operational decisions before these decisions are implemented should be a capability provided operational business intelligenceapplications.
- Operational business intelligence applications should enable operational business users to use the operational data, while simultaneously allowing them to make the changes to “staged data,” without changing the operational data.
Operational Business Intelligence
Wouldn’t operational users, such as inventory managers in the warehouses or truck drivers supplying goods to grocery stores, prefer having up-to-the-minute operational data? Wouldn’t this be better than the “old” data stored in data warehouses or data marts to make correct and rapid decisions immediately? Operational business intelligence is the solution.
Figure 2: Operational Business Intelligence
What facilities should be added to traditional business intelligence to make it an operational business intelligence application?
- Traditional business intelligence works with data warehouses or data marts. This data represents “old” data.
- Operational business intelligence provides access to “real-time” data as it happens to produce reports that can be used by operational users. With real time data, operational users can make decisions immediately.
- Operational business intelligence applications allow users to verify the ramifications of their decisions before they are implemented.
Because of this, we must go further than traditional business intelligence. There must be a facility where the operational users can make decisions based on real time data. They must also be able to verify that their desired changes (to the business processes) will provide them with the expected results.
- Operational business intelligence applications imply that the operational users can have access to interactive reports based on operational data. This is done so they can make split-second decisions about what should be done immediately. This capability allows the operational users to monitor the business activity as it happens.
- Operational business intelligence applications must be built on top of Online Transaction Processing (OLTP) applications to provide up-to-the-minute, and in some cases, up-to-the-split-second “read only” data and information for the operational users.
- Operational business intelligence applications should be able to use the operational data in a separate “staging” area, make suggested changes and play the “what if” scenario for the operational users in “real time.”
- The data supplied is not data sitting in a data warehouse or data mart from past years, months, weeks or even days. It is “live” data from the operational business applications, which represents the processes of the business in real time.
This is evident when we take real life examples from healthcare.
Traditional Business Intelligence
Based on nationwide data accumulated about heart disease over the past five years, there are now many recommendations. In terms of lowering the risk for heart attacks, it is recommended that people exercise moderately for about 30 minutes daily, eat 5 servings of vegetables daily (e.g. a serving consisting of 3 tablespoons of cooked vegetables) and eat 5 servings of fruit (e.g. a serving consisting of 3 tablespoons of pieces of fruit). This example reflects how past data can have an impact on future long term results.
Operational Business Intelligence
During a triple bypass heart-surgery, the surgeon must quickly decide when to perform certain surgical procedures. These procedures are not only important for a successful surgery, but also to save the patient’s life. Here, the surgeon is not thinking about whether or not the patient ate enough vegetables and fruit or about what the patient should do to avoid future heart attacks. At that time, the surgeon’s complete concentration is on the “immediate action” to save the patient.
Operational Business Intelligence Applications
Applications for future use can be written in a certain way that “sifts” through data about the habits of people from certain geographical areas. For example, heart disease happens more frequently in some parts of the country than others. The application is also written in a way that evaluates results and finds profiles of those in the “high-risk” category for heart disease. Furthermore, physicians are provided with the identities of people who could be victims of a heart attack before it occurs.
Figure 3: How operational business intelligence differs from the traditional business intelligence environment.
Operational business intelligence applications are built by professional application developers to support specific business processes. Characteristics of using operational business intelligence applications are:
- Operational business users, who are sometimes called "information consumers," should be able to make decisions quickly and decisively. Such decisions consistently affect their operational areas with a direct impact on corporate objectives and policies.
- Information consumers “in the trenches” should be making day-to-day, hour-to-hour and even minute-to-minute decisions with their areas of responsability.
- A business could come to a screeching halt if inventory control decisions are made by decision makers at the corporate headquarters. These people should not be making the inventory control decisions since they are so far away from the warehouses—both metaphorically and geographically.
- The high-level executives should not be making the minute-to-minute, hour-to-hour decisions regarding the operations of their enterprise. Instead, they should be asking themselves: “What business are we in? Which products or services should we be selling? Why are our sales better in the Northeast for the second quarter of the fiscal year than in the first quarter?”
- This is operational decision making. Business intelligence for this environment is much different than whether the trucks should be rerouted because hurricanes are hitting certain parts of the country. Rerouting the trucks is an “operational business intelligence decision.”
How do “Business Intelligence Suites” differ from “Operational Business Intelligence Applications”?
- “Operational business intelligence applications” provide both access to “real-time” data for the operational users and helps substantiate the decisions they want to make, which are based on the reports they receive. This means that “operational business intelligence applications” not only provide access to “real time” data, but also allows them to get access to “real-time decision making.” Most business intelligence suites work with data in data warehouses or data marts.
- The operational business intelligence application is an integral part of an Online Transaction Processing (OLTP) Application. An OLTP application should consist of two major components:
- Traditional OLTP Component: This component represents the automation of a business process or multiple business processes. An example of this is passengers making airline reservations. In this instance, data is collected about the passengers and flights. Various other activities are going on as well. Money is collected, tickets are written, the crew is scheduled, etc.
- Operational Business Intelligence Application Component: As passengers make reservations it becomes noticeable that certain routes are not doing very well on a specific day. The operational business intelligence application should determine an offer immediately while passengers are making online reservations. The offer could be a 10% discount if they reserve the flight in the next 10 minutes. This facility should be available to the airline reservation agents, or to passengers reserving their flights online.
- The operational business intelligence application (OBIA) component does not make any changes to the operational database. The OLTP component will make the changes to the operational database. In fact, the OBIA is developing strategies based on the real time data and logic programmed by application programmers.
- The operational users shouldn’t have to write application programs. We shouldn’t try to make programmers out of users. They should simply do what they do best—run the business itself. Trained application programmers are the people that should be writing the OBIA to co-exist with the business applications.
- The operational business intelligence application component could consist of two major parts. One part could be the use of the already programmed business logic by the application programmers, which is implemented automatically. The other part would consist of providing reports to the operational users, which would recommend the strategies to be implemented via the operational business intelligence application.
- Since OBIA focuses more on current happenings than future outcomes, the reporting facility of the business intelligence tool must be very comprehensive. The analytics implemented, though, could be light. The OBIA’s audience is more the operational users who must make decisions on the spot than the Power Users. Instead, Power Users consider the strategies that must be implemented by the organizations in the upcoming months and years.
Requirements for Operational Business Intelligence Applications
The user interface should be the same as the OLTP component of the application. The operational user shouldn’t have to learn multiple “look and feel” interfaces.
It also has to be interactive. The operational user should be able to put in the proposed change of action, such as the previously mentioned example of rerouting the trucks, and be able to see the impact of the change right away before it is implemented. This gives the operational user the capability to “tweak” the proposed change.
Where do we stand vis a vis “Operational Business Intelligence Applications” and traditional business intelligence vendors?
- Power User Audience: Most business intelligence vendors have concentrated on “Power Users,” supplying them with fancy tools to build cubes, dazzling graphs, zooming in and out charts. But the data they work with is data from data warehouses or data marts. They don’t use the data as it happens, but as it has already happened. The emphasis here is on analytics and trend detection. Reporting is only used to publish the results. However, there is no impact on what is happening now. Traditional business intelligence vendors have not included reporting with “real time” data in their offerings, let alone interactive reporting.
- User Interfaces: User interfaces for the business applications and the applications supporting business intelligence have major differences. Consequently, the operational user that is using the business application feels alienated.
- Operational Business Intelligence Applications: The application should support interactive reporting and simple analytics. Interactive reporting with the operational data, including the ability to verify the results of your actions, is the key difference between traditional business intelligence vendor tools and the operational business intelligence application development platform vendor.
How to make the Operational Business Intelligence Architecture Feasible?
- The biggest problem with operational business intelligence is what happens if you get it right. Scalability is especially important.
- The operational business intelligence application environment is not a replacement of the traditional business intelligence environment. It should be complimentary to the traditional environment.
- The future is happening now for operational business intelligence applications by providing “real time,” “read only” access to the transactional data. This should be accompanied by comprehensive reporting capabilities and simple analytics.
- After seeing the “real-time” data in their reports, operational users should be able to provide changes within the operations of their functional areas. The OBIA component, which accompanies the OLTP component, should provide a facility for “what if” scenarios for the operations. Then, operational users can be “proactive,” rather than “reactive.” And isn’t that what business intelligence is supposed to do anyways?