Analysis & Commentary:

BUSINESS INTELLIGENCE - CHOOSING THE BEST SOLUTION
by Petr Havlik, MD, Cyberpro

The local corporate market today is full of business intelligence (BI) solutions ranging from comprehensive suites, to middle and small business BI tools. Question is what BI solution does one choose?

The market is brimming with BI solutions ranging from comprehensive suites such as integrated Enterprise Resource Planning BI tools, Cognos, Crystal Decisions, SAS solutions and Oracle solutions, to middle of the road products such as Crystal Reports right through to makeshift small business BI tools such as MS Access and Excel.

BI is the field of corporate data storage, organization, dissemination and reporting in order to enhance the operation of a business. Since BI is such a diverse subject with infinite variables, it is essential to either employ an expert or consult external experts while choosing a BI solution.

Points to consider when choosing a BI solution:

Technical compatibility:

Far too often sales people selling BI solutions will brush over technical issues, stating that technology is just an enabler and that the real issue is the business. This is all fine and well, but at the end of the day the technology has to work. All too often customers find out half way through an implementation that certain requirements can't be met due to technical issues. Ensure that the technology you are implementing is in line with existing and planned IT infrastructure.

Simple tools are not effective at operating in multi-user and multi-location environments and a tool that can support this must be chosen. Specific tools are designed to plug into specific ERP and line-of-business platforms and this must be considered when looking at a BI solution.

What are the business requirements?

BI tools offer functionality ranging from simple reports, drill-down capabilities and On-Line Analytical Processing (OLAP) functionality to solutions targeted at specific industries and operational environments. Many tools allow end-users to design their own reports and handle multi-user and access control considerations.

The choice of tool will depend on whether a simple reporting solution is required, or more in depth functionality is necessary. For businesses running ERP systems, most ERP vendors offer some form of built-in BI functionality in their products. Sometimes this is in partnership with third-party BI vendors, but is typically sold as a unified solution. If this functionality covers the business's requirements then it is probably the best bet as there is a trend towards integrated solutions that do not suffer from incompatibility problems and reliance on multiple vendors.

Often a business will find that it needs to develop a custom solution due to cost or functionality constraints presented by an off-the-shelf solution. This is advisable only if the requirements or the financial considerations really demand it, as a custom solution will always suffer the risk of becoming unsustainable or obsolete due to its proprietary nature. On the other hand, if a good development partner is chosen a proprietary solution can sometimes offer far superior functionality at reduced costs.

What budget is available?

Costs of BI solutions can range from a couple of thousand to millions of rands. One must consider the following costs:

• Upfront package costs.
• Annual or other ongoing licensing costs, if any.
• Per-user licence costs.
• The costs of hardware and infrastructure required to run a solution.
• The costs of consultants and developers who will be required to implement the solution and maintain it in production.
• Costs of tying up internal resources in creating a BI solution.

What existing BI solutions and IT infrastructure are in place?

Ideally, a BI solution needs to leverage of existing IT and BI investments to ensure that previous IT expenditure does not go to waste and to ensure technical compatibility. The exception to this rule is if previous infrastructure implemented is clearly insufficient to support an effective BI solution. Try to make sure that the BI solution implemented will be able to accommodate historical data as this existing data is extremely valuable to the business.

Systems that BI relies on:

Enterprise and e-commerce system databases, documentation, intellectual capital and third-party information. These are the sources of information that are effectively the reason for a BI need. These sources of information are often very diverse ranging from ERP systems such as SAP, Baan, PeopleSoft, Oracle and various others to data stored on managers' PCs in Excel and Word documents.

Line of business and complementary applications used by the business. These are the applications that the business uses to capture and manipulate data in the process of doing day-to-day business. Some examples of these are point-of-sale systems, customer relationship management (CRM) systems and accounting systems. These applications all store large amounts of useful data, but without a way to access this data in a concise and summarised way, the data is useless from a business decisions point of view.

The components of a BI solution

A typical, high-end BI solution will consist of the components below. These are often included in many high-end ERP systems, provided by BI solution vendors or developed in-house.

Data warehouses and data marts. The role of these is to combine all the relevant information from disparate corporate systems and data sources. This process involves the transformation and vetting of data that is required for the BI solution. Data is typically stored in a translated format to make it suitable for high speed reporting requirements.

A data warehouse is a large collection of all required corporate data. Sometimes there is a requirement to split this data into smaller functional units. These are known as data marts. Data is split to accommodate differing requirements for different business units, to speed up the reporting process, for technical reasons and sometimes to allow faster implementation of specific requirements. At the end of the day a data warehouse or a data mart is a glorified database organised in a structure appropriate for reporting purposes and containing all relevant data required for reporting purposes.

ETL (Extraction, Transformation and Loading) software. This software is responsible for gathering data from the various data storage systems in the enterprise, transforming it, checking it and loading it into the data warehouse or data mart. This is an ongoing process to ensure up-to-date data is available for reporting.

This piece of the BI puzzle has a complex task to perform, as it is sometimes extremely difficult to deal with disparate, dirty data from numerous systems. Dirty data, that is data that has not been captured correctly or completely via the CRM or line-of-business application, can mean the success or failure of a BI solution.

Report-generating and presentation tools. These are the actual tools that allow either IT personnel or decision-makers to create the required reports. Most off-the-shelf BI solutions come with comprehensive report-building tools, but businesses will often find that they need to develop custom solutions or add-ons to achieve the desired results.

Report presentation tools are the tools used by the intended audience to view reports. The current trend is to create Web-based reports that allow easy access from remote offices. Tools will almost always take security considerations into account, ensuring that sensitive information is only accessible by the appropriate decision-makers. Tools are also able to drill down into data as well as slice and dice reports in various ways giving decision-makers the power to examine data of interest in detail. OLAP tools give users the ability to examine multiple data dimensions, known as cubes, allowing the user the power to look at a number of variables at the same time.

The benefits of a well-implemented BI solution in today's business environment cannot be doubted. A business thinking of implementing BI needs to assess the internal information requirements, look at whether these are being met and at how much overhead in operations, as well as how much competitive advantage, profit and customer understanding is lost by not having a BI solution in place. The jobs of creating the appropriate data warehouse structures and initial reports typically fall on an external data warehousing consulting firm, but require strong involvement from internal business as well as technical parties to succeed.