By Abhishek Kundan, Product Management & Strategy Consultant, Fintellix Solutions (formerly iCreate Software)
Banks produce copious amounts of data daily through customer transactions, financial activities and from external markets. The raw data in itself has little value unless it is properly captured, consolidated, stored and extracted into ‘actionable information’. This information is further distributed, analysed and presented into actionable insights for driving business decisions. Here’s where Business Intelligence (BI) tools come in to leverage the data to extract maximum value.
There’s no dearth of options available for a bank to leverage data with BI and there are quite a few vendors with variety of BI products and services, with interesting features and functionalities. There’s also the option to develop in-house solutions or go the conventional spreadsheet route for data analysis and reporting. Success however really hinges selecting, and more importantly effectively using the ideal BI approach based on the bank’s size and needs.
A serious introspection therefore would be in perfect order before going in for the big BI decision –
- How flexible is my BI tool’s Data Management capabilities? How to convert raw data into BI-usable data? Is my BI tool scalable and fast enough to absorb a sudden data deluge? And what about data quality?
- How extensible is my BI process to new technology or new module add-ons? Will it integrate well with other analytical or statistical applications?
- How great is my BI tool in terms of user experience and visualization? How business-user-friendly is my BI process? Will they understand technical terminology?
- How solid is my overall BI framework and process in terms of security, administration, control and accountability?
Let’s attempt answering these questions and discovering the elusive ‘right approach’ to BI for better decision enablement in banks.
BI in the Banking context.
BI is not a new concept to the banking industry and application has evolved radically from manual systems in the pre-computerization era to modern sophisticated BI platforms. The archaic process of BI was limited to management and regulatory reporting; and involved cumbersome manual consolidation of reports across branches.
With rapid technological advancements and market expansion, transaction processing and data aggregation got automated. This helped manage voluminous data and reduced the risk of errors while handling data manually. Banks eventually transitioned to the more structured and sophisticated MIS (Management Information System) for their BI needs and aid the decision making process.
Conventional approaches to BI.
Often, BI tools are deployed directly over the source systems for data analysis. This is suitable for small-sized banks which have lesser branches or limited geographical presence and relatively smaller size of operations. This plain vanilla BI approach may result in faster implementation, but carries multiple drawbacks. Since the BI tool is directly deployed over the source system, it provides a departmental view of data, rather than at an enterprise-level.
Also, this approach doesn’t allow cross-functional or multi-dimensional analyses of data, thus giving limited or incomplete insights to the business. Business users need to collect data from various information silos depending upon the type of data they want and then consolidate manually to get insights for decision making, which consumes unimaginably substantial amount of time, efforts and resources.
A more pervasive approach is Enterprise Business Intelligence, where data from multiple source systems across the bank is extracted, transformed and loaded into a Data Warehouse (DW) for analysis. Various BI applications like reporting, dashboards, data mining, analytics, etc. are added on top of the DW as per end user requirements. The advantage of having an integrated BI architecture is that data can be pulled from multiple disparate source systems and a wide variety of cross-functional analysis and multi-dimensional analysis can be performed based on business user needs. The output can be in the form of operational reports and Departmental Dashboards for Middle Management, Executive Dashboards for Senior Management, Regulatory Reports for Compliance, Business Performance Management for Business Users and so on.
While a DW serves the needs of BI at an enterprise level, a Data Mart which is a subset of DW, serves specific needs of specific business functions, e.g. a Risk Management specific Data Mart. It also helps optimise the performance of queries and delivers results faster as queries traverse through a limited data set rather instead of the entire data in the DW.
Challenges of existing BI process.
However, applying the Enterprise BI approach has its share of challenges. From Core Banking to Loan Origination to Treasury Management to CRM to Wealth Management, a bank today can have over a hundred source systems for storing diverse data. All these systems are designed differently and implemented differently and therefore may not share a common structure. This presents a big challenge in integrating and managing data under one roof, but parked in disparate sources.
Another obstacle is that of integrating data absent in the source system to the Warehouse. Ex-source system data is mostly unstructured and is usually present in external files like isolated spreadsheets, word documents or log data.
With dynamic regulations, fiercely competitive business environments and a fastidious customer base, changes in the global banking sector are now occurring in an exponential fashion. Needless to say users in banks simply must have relevant, accurate information at their fingertips to take the best decisions. Vital therefore to deliver the right insights to the right person at the right time. Equally important is information delivery, in terms of visualization and experience.
From front-line personnel to senior executives with specific data needs, there are a number of BI system users at different hierarchy levels. Given the accessibility to data of strategic importance, it presents a problem in managing security and control over the BI system. An unauthorized access to sensitive information can pose serious threats.
A smarter approach to BI.
The tribulations of conventional BI systems can be converted into opportunities by making a more robust, enhanced version of the system with a common platform foundation for utilizing data effectively and enabling ongoing maintenance. This approach features the addition of a ‘foundation layer’ that supports the data flow end-to-end right from the source systems through to the output (as reports or dashboards).
A platform provides a consistent, convenient and complete interface for consolidating, enriching and exploring data with consistent business terminologies and logic, making life easier for business users. A good BI platform has the capability to integrate with other BI solutions and allow additions of new custom modules, making it more extensible and flexible. Functionalities such as trigger-based alert mechanisms or communication features like sharing & commenting can all be built over the platform, creating more visibility and collaboration, eventually improving efficiency and productivity.
So what differentiates a platform approach from conventional options? A good platform-oriented BI system should essentially have four key dimensions.
Data management.
A BI platform should provide a standardized interface for disparate source system for data integration with user friendly banking nomenclature. It should have a provision to upload external data such as spreadsheets or text files into the BI system. It should also have provisions to import/export missing data to capture gap data. For example, data related to branch information is typically not available in standard source systems and is maintained manually using spreadsheets. An adjustment facility either at an account level or at report level, with proper justifications and audit trail mechanism, would be advantageous.
Data processing.
The process of data discovery has to be simple and user-friendly (no complex technical codes) with absolutely minimal support from IT. The system should provide the ability to trace back data to its origin and drill it down to a granular level. The platform should have the ability to support and integrate with statistical engines (such as SPSS or SAS) for advanced analysis and predictive analytics.
Different users have different needs for the way they want data. While a business analyst may need data for sales or product analysis as a spreadsheet, senior executives may want it as graphical charts to get a high level view of business unit performance in a portable document format. On the other hand regulators may require report submissions in XBRL so that it is compliant with industry standard formats. The BI platform must provision for supporting multiple formats or smooth transmission via e-mail.
Security, control and accountability.
Given the number of users accessing the BI system, preventing unauthorised data access and alteration is another key consideration. A multi-layered security framework with role-based data access in line with the bank’s internal information management security system, should be inherent in the BI platform. Data security should be applied across the bank’s hierarchy depending on the business requirements.
A complete and historical audit trail of all changes made in the BI system should be readily available to track precise details of the change – who, when, what. System usage tracking and activity monitoring tools also aid in effective administration and management of the system. Moreover, workflow management tools such as Review process flow can be built over the platform to track accountability of any task and it also helps in tracking delays and monitor status through alert based mechanism.
User experience and collaboration.
User experience is much more than just the ‘look and feel’ of the product; it’s about how the system actually ‘connects’ and interacts with users. A BI tool is meant for taking better decisions for business, thereby increasing productivity and efficiency. But a complex and clunky tool can be detrimental because of its design and processes. A good BI system is simple, intuitive and easy to use.
Functionalities such as Self-service BI, whereby reports and charts can be generated on the fly, or trigger-based alert mechanisms or sharing and commenting features, all help in optimising experience while enhancing collaboration.
While there are multiple BI options available to banks, the decision to choose a platform-oriented approach over a traditional one is determined by a bank’s size, scale and strategic direction. Most banks concentrate on flexibility of a BI tool to accommodate all aspects of Compliance, Risk and Analytics, and the extendibility of the tool to accommodate future requirements (statistical analysis for instance) as the primary differentiating factors. But, disregarding the platform aspect would be short-sighted and may result in issues such as data quality, user reluctance, security breaches and sub-optimal performance processes. A long term strategic investment in BI therefore should not just concentrate on flexibility and extendibility, but also seriously consider the platform dimension.
2 comments
Brilliantly written and absolutely useful. Thanks for sharing.
Comprehensive and well written but that it is a bit too long