Investing in better understanding

Financial institutions were among the first users of BI, predominantly for analysis of customers and risk. In this respect, they are considered very advanced in terms of their BI implementations – in many cases using complex data mining tools to better understand their customers.

However, according to Richard Brown, Head of BI at SAP, there is typically a lot of data within these organizations’ operational systems that is still left untapped. “Once you take away the customer analysis piece, financial service organizations are like many other organizations; they still have to deal with sales, purchasing, finance, marketing and human resources,” he says. Understanding how these operate and perform is therefore critical in decision-making throughout the business.

Simon Cattlin, Regional Director for Sybase, stresses that business intelligence is largely focused on historically structured reference data. “The ability to effectively search through this type of information is vital for audit and compliance reasons.” However, “80 percent of data in these organizations is unstructured – for example e-mail and CRM voice files – and is often held in disparate silos.”

As such, the challenge for financial service organizations is how to get the best value and performance across the whole data set, linking real customer information with reference data, together with reporting on regional lines and time lines. This is where BI comes to the fore.

Applying BI

Andy Hirst, Director of Worldwide Financial Services Marketing at Business Objects, believes there are four major areas where financial services organizations are applying BI to improve performance. First, he believes that BI is widely used to “mitigate risk and for complying with Basel II.” It is used to report from credit risk engines and display key risk indicators to improve capital allocation. He points out that all manner of risk reports – whether by counter-party, industry or product line – can be automatically produced and displayed in reports or risk dashboards.

Second, BI is being used for customer insight. “For example, our customers are using BI to identify cross-sell and up-sell opportunities in both retail and wholesale banks,” he says. This improved access to customer information empowers employees to be able to cross-sell new products and services to existing customers. By implementing BI, they can use management dashboards to track and monitor sales and product metrics; regions that are above or below forecast in sales and product categories; and line-of-business leaders in mortgage or small business lines.

Third, Hirst points out how BI is used for real-time financial reporting and planning. “A huge growth area is using BI to improve net interest margin spread through improved planning and cash flow management.” Using business intelligence, financial services institutions can model the impact of new branch openings and the schedule of capital expenditures to optimize financial performance.

Finally, says Hirst, business intelligence is “embedded into a lot of banking operational systems, such as Iflex and PeopleSoft (PSFT) products. Business intelligence is used in all manner of operations to report on product management information on the operational process – for example, HR reporting (distributing compensation reports across the company from operational PSFT systems); ATM reporting (usage and uptime reports on the ATM network); IT cost allocation reporting (improved allocation of reports on IT usage); and core bank reporting (key operational metrics such as number of new applications or transactions completed by business areas).”

Best practices for better ROI

Does this application of BI therefore bring in the returns that are so important to the bottom line, and how can business intelligence result in a positive ROI? Hirst believes there are three key ‘best practice’ guidelines organizations should adopt to ensure maximum ROI form their BI deployments. “First, they need to produce clear, actionable management information early in the project (within three months) to demonstrate value to the business,” he says. “Second, they must empower employees to run their own queries so they can get the information they need without recourse to IT. Third, BI is a powerful tool that can be used to improve every area of the business, so they need to train employees on how to get the most from BI.”

For SAP, the key to achieving ROI from a business intelligence investment is making the information available ‘in-place’, or where the business decision is being made. “Often, BI is an after-the-event analysis tool looking at ‘how we got it wrong’. But SAP is driving to embed the analytical information in the business process applications so that when people need to make a decision they have all the relevant information to hand,” confirms Brown. “And we are not just talking about the strategic decision-makers; the ROI is delivered by empowering the foot soldiers to make informed decisions, too.”

This may well be the case for some, but Cattlin raises the point that ROI is, for many banks, a relative issue. “Most banks tend to measure themselves on cash utilization ratios, which have a different economic impact to ROI. As such, ROI is not the main concern to a financial service organization investing in a new data management and BI system. The main challenge is to provide a superior service to increasingly mobile customers, whilst meeting compliance regulations.”

Cattlin’s point strikes upon another major consideration – that of regulatory compliance. So what are financial institutions up against? “Looking at the regulatory environment, there are well over 50 types of regulation that can affect an organization’s license to operate,” continues Cattlin. “Financial service organizations have to reply to mandatory audit checks, which also apply to unstructured data such as e-mail disclosures, both internal and external. They have to rethink the way they store this type of data so that it can be analyzed in such a way that it is auditable and controllable, making it easier to detect errors and prevent fraud. And they must be able to demonstrate and control repeatable processes. Many banks are around two years behind the regulations and are reactively compliant. The current focus is for them to become pre-emptively compliant, where they will be using analytics to maintain a competitive advantage.”

The key drivers of SOX, for example, are compliance and transparency within an organization. Information technology can do a lot to achieve these objectives by providing auditable standardized business transactions that can be procedurally controlled. “SAP has gone on to extend these transactions with tools that look to audit the procedures and look for conflicts of interest. A whole range of standard checks are made to ensure the segregation of duties, but to this can be added any specific checks that an organization deems necessary,” says Brown.

Identifying these is only half the battle, however, as Brown points out. “Like any other piece of management data, we need to understand the trends: the ‘where and how’ of which areas have the highest degrees of conflict and how quickly these are being resolved. And more importantly, when they are resolved, we need to have monitors to check if they will recur.”

Analytics, and a robust BI framework linked to the core transactional systems, also ensure the ability to deliver clear and unambiguous financial information for compliance reporting. The credibility of this information is based on the data being collected systematically with no ‘manual massaging’ and with clear drill-back to the detailed transaction record. The framework must be flexible enough to change over time as new ways of beating the system are sought. This is as much a part of delivering SOX compliance as any other aspect.

Furthermore, if data is not effectively monitored and managed, costs and inefficiencies can escalate. Cattlin believes data must be gathered from the disparate systems it is held on and held efficiently in a single, large-scale warehouse structure for maximum consistency. “Data can then be analyzed centrally for an overarching view of the whole business,” he says. “Companies need to take a holistic view of their data in order to improve BI. To be effective, this must be managed as seamlessly and cost-effectively as possible.” The driving force behind this decision must therefore come from senior management, as it requires a holistic view of the organization and cannot be driven effectively at departmental level.

Greater attention to architecture

Many software organizations are re-architecting their products to have a service-oriented architecture (SOA). Phil Wood, Product and Solutions Marketing Manager at Business Objects, defines SOA as an approach that segments and isolates application functionality into smaller, discrete and usable components, otherwise known as ‘services’. “The primary goal of an SOA,” he asserts, “is exposing application functions in a way that is reusable and standardized so that once created, they can be leveraged across multiple projects.”

Taking a similar approach to the design and deployment of an enterprise BI solution can have similar benefits. “Most organizations are seeking to reduce the number of independent software applications deployed: driving enhanced supplier relationships, reducing latency in the information supply chain and ultimately making the organization more competitive in it’s chosen markets. As the rate at which organizations grow, acquire and divest is only going to increase, it is vital that the underlying software infrastructure is equally dynamic and able to adapt to business change while continuing to drive cost from the IT infrastructure,” says Wood.

Brown continues: “BI solutions are not like transaction processing systems. BI can often be a tactical, subject-specific solution that does not need to be applied across the organization in a ‘big bang’.” As such, developments can be very subject-specific (or flavored). This is often not an inhibitor at first, but as the demand for BI grows so we see the same data being extracted many times with different purposes and therefore different ‘flavors’. “This will result in conflicting information. It also means any new requirement needs to be a subset of an existing data mart or needs the data to be extracted once again.

“At some point in any BI development, there will come a time when some fundamental architecture will need to be laid down – a data layer to support all BI requirements. This layer needs to be detailed, integrated and unflavored. Where we have seen organizations that are able to do this, we also see that they are able to deal with new and changing BI requirements swiftly and effectively.”