Failure is not an option

by Melanie Liew

December 24, 2003 —

The key to working smarter involves developing a deep understanding of customers to enable cross-selling and customer retention.

In fact, these days, the true differentiator for any company trying to separate itself from the pack is its ability to construct a single view of each customer. Using puzzle pieces obtained at each point of customer contact, the intelligent enterprise can predict how customers will respond to offers and determine how to focus organizational resources to deliver greater customer value.

Paul Castle, managing director, SAS Institute Inc, said, "Today, employees need to understand the drivers of success in all facets of the organization, anticipate the impact of changes in key variables and measure progress toward both divisional and enterprise objectives. This will enable them to consistently deliver value to customers."

In fact, a broad range of applications for BI is enabling organizations to garner a high return on investment figures. BI is used to identify cost-cutting ideas, uncover new business opportunities, roll ERP (enterprise resource planning) data into accessible reports as well as react quickly to retail demand and optimize prices.

Castle added, "With traditional applications, companies can improve the efficiency of operations, for example, processing transactions more efficiently with less manual intervention or improving communications between departments within the company. But these operational systems such as enterprise resource planning systems, sales force automation systems, customer interaction systems and procurement systems are not capable of creating the knowledge that a firm needs to differentiate itself and to sustain competitive advantage."

For this reason, BI is increasingly recognized as key for business growth. According to research company IDC, the BI market is projected to grow to US$7.5 billion in 2006.

But, the situation is far from being cut and dried.

Kriss Channe, marketing director at Business Objects said, "Few organizations have a comprehensive BI strategy or clearly defined BI standards, and so face a growing patchwork of disparate BI technologies."

A big challenge that companies encounter, said Castle, is the proliferation of disconnected information silos. Currently, individual departments have been run like separate businesses, each free to pursue its own IT infrastructure.

In fact, Moaiyad Hoosenally, industry manager, Technology Practice at analyst firm, Frost & Sullivan said, "Our research indicates that while multi-channels have been rolled out, few organizations have managed to develop a consistent customer experience across the channels or dramatically reduce the cycle time from inquiry to fulfillment. This is partly because workflow systems have been inadequately implemented."

This has resulted in the use of disparate hardware, platforms, systems and databases throughout the enterprise. In addition, it becomes extremely difficult to establish and measure progress toward company-wide objectives.

"Companies are using BI to justify or disprove the wisdom of what would otherwise be gut business decisions."

"Without BI tools, very often, the evaluations of opportunities for growth are based on gut feelings, 'guess-timations' and assumptions because it would be too expensive and time consuming to get hard data. BI can let organizations run some quick and credible numbers to justify that gut," said George Lee, managing director, ASEAN & India, Hyperion Solutions.

Another factor that companies have to deal with is the convergence of the different tools within the firm. Said Forrest Palmer, managing director, Asia, Cognos Far East, "There are so many different report types in companies today that the line between the different types of reports is blurred and continues to blur more and more. Production reports (invoicing), management reports (monthly sales), analytical reports (a comparison over time) and ad hoc reports have typically been done with different tool sets."

"This means users may have one tool to do their ad hoc reporting, another tool (with a different interface) to call up their standard reports and yet another tool if they want to write their own reports. The IT department probably has another set of tools to do their professional report writing. This means different metadata, different calculations, different names for the same business entities and all this leads to different results and a high cost of maintenance. It also slows down responsiveness of the organization to rapidly changing business needs."

And, there is no single version of the truth.

Therefore, the situation has created an increasing demand for solutions that can overcome these obstacles.

To be sure, the stakes are high because the cost of incomplete, inefficiently generated intelligence is astronomical, not only in missed revenue opportunities, but also in real cash outlays. Said SAS Institute's Castle, "Organizations have wasted large sums on niche software products that are expensive to integrate and maintain, do not provide one version of the truth and leave decision makers in doubt of their output."

Also, more recently, many BI companies are consolidating. Said Palmer of Cognos, "Every time a company acquires another, not only do they get their products, their financial surpluses or deficits, they also get a new IT infrastructure which usually includes a whole new set of BI tools. Similarly, vendors in the BI space have also been consolidating. This means that each time a BI vendor gets consumed, the existing customers then have the task of rebuilding their BI applications in the new tool or else they have to re-evaluate their options."

So, even with BI solutions, it is important to standardize the tools within the enterprise. Standardization will mean that users within the organization need only learn to interact with one product. For IT administrators, there is an elimination of redundant efforts on installation, configuration, tuning and load balancing as well as maintenance and patching.

But, even before embarking on a BI project, IT administrators should make a business case for standardization. The project owners should define the business requirements, define the scope and intended impact by standardization on the BI tools across the organization.

Key to the project, is the need to take inventory of what is being used and what it is being used for. Often, firms buy many tools for one purpose and are not used beyond that capability even if they are capable of much more.

And, businesses must assess the functional capabilities the users need. For investment protection, organizations should address the needs of the company in the future.

Whatever it is, take the integrated approach to building a data warehouse from the beginning. Organizations should make sure that they are not locking themselves into an unworkable data strategy further down the road.

The plan should also include costs, resources, milestones and dates.

Companies should assess carefully, the vendors that can meet the requirements and select the BI solution that provides the best fit.

Cautioned Palmer, "This should not be undertaken if the executive team does not see any value from the project."

Once buy-in from top management is secured, the project needs the sponsorship of a senior executive without which, the project will not succeed.

Said Lee from Hyperion, "Companies that embark on a BI project should deploy the project quickly, then adjust as it goes along. Don't spend a huge amount of time up front developing the perfect report. This is because needs will evolve as the business changes. Deliver reports that provide the most value quickly and then tweak them."

Jim Koerner, vice president, DB2 Information Management, Software Group, IBM Asia Pacific said, "The scope of the BI architecture has expanded beyond the traditional core warehousing disciplines such as data extract, transform and load (ETL) and reporting. Now, BI requires information on demand, that is, actionable business insights combining data mining and multi-dimensional data analysis with advanced statistical and analytical functions in a real-time integrated environment."

OLAP (online analytical processing) and mining, in particular have brought with them specialized tools. These include application programming interfaces, data structures and sometimes even purpose-built engines spanning all levels of the system stack, from front-end client tool to middle-tier server cache to backend data warehouse. The multi-tier, enterprise-wide nature of BI tools and infrastructure demands a high-level architectural framework for BI.

For example, IBM makes the BI function in the database accessible solely through open, standard interfaces as part of an integrated BI platform and collaborate with partners for other layers of the architecture.

IBM DB2 Universal Database supports BI functions inside the database. These BI capabilities include data mining, OLAP, ETL, spatial and advanced statistical and analytical functions for regression, covariance, sampling, ranking and moving windows. OLAP tools are very helpful to analyze the business. For example, the effectiveness of a marketing campaign depends on many variables and a pre-defined report might not show the right information.

Said IBM's Koerner, "Many BI applications use simple SQL syntax to extract data from the warehouse in bulk, stage it within the application tier and filter or aggregate it to produce the desired granularity. Only then do they apply the functionality. That is their value add."

That said, Castle of SAS Institute added, "Intelligence does not depreciate in value. It is not diminished by bear markets or global recession. In fact, as economic conditions grow more precarious, the value of intelligence increases since it is the only resource that enables an organization to adapt its strategy quickly to meet any challenge."

Lee concurred. He advised going for the payoff first and added, "Keep it simple. Guard against the tendency to do something fancy. It is better to build a simple infrastructure, get it into the users' hands and let users tell you what they need. At the end of the day, don't deliver reports," he said. "Deliver tools."

"Too many executives today are in survival mode and not investing in adding value. The irony of this is that by not investing, executives may threaten their companies' chances of survival."

Computerworld Singapore