Operational BI for Banking
Operational BI embeds analytical processes within the operational business structure to support near real-time decision making and collaboration. This characteristic fundamentally changes the way how data is used, where it exists and how it is accessed. This change is rapidly exposing the limitations of traditional analytical tools.
Operational BI helps businesses make more informed decisions and take more effective action in their daily business operations. It can be valuable in many areas of the business, including reducing fraud, decreasing loan processing times, and optimising pricing.
Caters to middle management and frontline: Operational BI delivers information and insights to those managers that are involved in operational or transactional processes.
Customer service executives in branch, get a flash on their screens on the likely requirements of a customer (based on his profile and past transaction behaviour) while servicing a customer request over the counter. The executive at the counter can use this information to cross-sell other products to the customer, during the brief period the customer was at the counter. This is an example of operational BI being used to collect and analyse information midstream-running BI tools directly on transaction systems or using enterprise information integration tools to query the data where it lies. The ability to get a glimpse of the entire pipeline in and outside of the company is very valuable to the agility of a corporation. A loan recovery manager would do well to tailor the language of the demand notice or the phone call based on the total relationship of the customer (and may be his family) with the bank.
Just-in-time delivery: The information needs to be delivered in near real-time (within minutes or hours) for the purpose of managing or optimising operational or time-sensitive business processes.
The objective of operational BI is to reduce the time it takes for a line of business user or application to react to a business issue or requirement. This elapsed time is known as ‘action time’. The business case for operational BI is based on identifying business situations where reduced action times can bring business benefits to the organisation.
For instance, the ability to detect and react more quickly to the fraudulent use of a credit card is a good example of how operational BI can provide business value. By analysing the history of fraudulent situations, the BI system can be used to develop business rules that signify potential fraud, and operational BI can be used to apply those rules during daily business operations. The closer to real time the fraud can be detected, the less is the operational risk.
However, not all operational BI systems need to be near real-time. Reducing action times to close to zero are is beneficial only in specific types of business requirements such as the fraud example.
In fact, operational BI can be classified into being demand-driven and event-driven, the latter being more automated. If the action time requirement is a few hours, business users or applications can use the BI system at on-demand analysis and evaluate the results manually to determine whether any action is required. In the demand-driven case, it is the user who drives the BI system.
But if the action time requirement is two seconds, then on-demand will not be suitable. In this scenario BI systems must track business operations continuously and automatically run analyses to determine whether any action is required. If it is, the business user must be alerted about the situation and sent recommendations on potential courses of action. In case of a fraudulent credit card transaction, the BI system is expected to refuse authorisation. In event-driven BI, business operations and the BI system drive the user. It is obvious that the implementation of event driven operational BI is more complex than demand-driven BI.
Uses recent transaction data: Data used for operational analysis is frequently accessed before getting loaded into the data warehouse. The latency in a traditional data warehouse implementation results from the batch mode in which it is populated. It is more suited for strategic applications such as historical analysis, risk management, performance management etc. But a dashboard needs to be as close to transaction data as technically feasible.
Less aggregation, more granularity: In a sharp contrast to traditional BI in which pre-aggregation, with optional drill down to detail levels is a norm, operational BI normally requires more of data granularity to address the needs of the specific operational function it supports. Traditional BI aims at a holistic view of corporate performance, while operational BI is process and user specific. Yet, some operational BI requirements do require aggregated data, such as the lifetime value of a customer, which is required for a directed sales call.
Embedded into business processes: Operational BI is intricately connected to transactional business processes. The extent of this integration depends on the level of implementation. One could use it to generate operational reports to analyse processes, or monitor them using dashboards and scorecards. In these two levels there is not much of integration. In the other two levels, where operation BI is embedded into business processes either to facilitate them (demand-driven) or to execute other processes (event-driven), it is embedded into the process.
Handles disparate sources and unstructured data: Traditional databases and data warehouses do not take into consideration the increasing use of unstructured data; such as emails, telephone calls, letters, internal notes etc, stored outside these systems, which are of critical value in an operational BI implementation. Another issue that it has to handle arises out of the disparate transaction systems in use in most of the banks. The variety of banking services makes it very complex and often impractical for a single software solution to handle all kinds of transactions. Extracting data from such disparate systems and making use of unstructured data is required to be handled by an operational BI system.
Availability is a concern: The high level of integration with transactional business processes demand the same level of availability from operational BI implementations that transaction processing systems have to provide. An outage of an operational BI application could have a direct impact on the organization’s ability to do business or to service its customers. Therefore, availability becomes a critical issue for operational BI applications.
Requires different architecture: Traditional BI vendors had built their products using proprietary architectures. While these architectures are ideal for strategic BI, they are not suited for operational BI. Because operational BI entails coupling BI applications with operation applications and operational processes, a component-based, service-oriented architecture (SOA) is necessary to fully support operational BI. Service-oriented architecture that lets users access real-time knowledge with a set of service feeds can maximize business agility while reducing complexity. For example, SOA flexibly and cost-effectively supports the midstream, on-the-fly data collection and analysis necessary for operational BI. Service orientation also supports operational BI throughout the business by pushing BI data out to the mobile workforce and enabling workers across the enterprise to incorporate this vital data into their workflow. The straight-through processing requirements in the banking industry necessitate immediate risk analysis, which in turn requires an online BI capability.
Posted on May 11th, 2009 by Ashwin Dedhia
Filed under: Business Intelligence, Emerging Trends
While it is correct that BI is much helpful at operational levels as well, it needs to be seen as to what is to be called as BI and how it differentiates itself from DATA INTEGRATION reports.
One of the points highlighted above was on the DATA LATENCY or ACTIVE BI.
Traditionally, BI focusses more on the repot delivery.
And for the purposes of MINIMAL LATENCY for reporting, the thrust is more on innovative kind of technologies.
Virtual data warehouse, Data federation are the current active basis which enable real-time BI.
I had no idea BI can be so useful in Banking industry. Very glad you have brought this out into the open so everyone in the industry can be aware and benefit of this.