Ten things you need to know about business intelligence

Driven by the quest for improved information quality, regulatory mandates, and the need for an enterprise-level view of the business, companies continue to invest heavily in business intelligence (BI). Aubrey van Aswegen, MD of Knowledge Integration Dynamics (KID), looks at the top 10 trends shaping the BI world today.

Gartner reports that BI is number two on the top 10 list of priorities for CIOs. Other research companies reported similar findings.

The fact is companies are making huge strides in BI as their approach to the technology matures. However, there are several key aspects companies need to be aware of to derive maximum value from their BI investment.

Compliance is forcing companies to take action on data quality, a complex concept that encompasses many data management techniques and business quality practices, applied iteratively, to achieve required levels of quality.

Data quality includes technology and business practices that improve data, such as name-and-address cleansing, record matching and merging, housekeeping, deduplication, standardisation, and appending third-party data.

Systematic measurement, iterative quality improvement, and verification are key components of a data quality programme. Data quality requires close collaboration among IT and business professionals who understand the data and its business purpose.

The need for a "single version of the truth" has underscored the need for master data management (MDM). Master data describes an organisation's customers, products and vendors, with the goal of ensuring semantic consistency across organisational and business process lines, and simplifying process and data integration.

This type of data, however, is typically managed by disparate, redundant and often external information systems. Not addressing the increase in complexity and redundancy of critical enterprise data can result in significant impacts across the value chain, such as delays in time to market, reduced productivity, higher supply chain costs and decreased customer satisfaction.

Master data is not as much about technology as it is about securing business-side involvement and ownership in the process. And it's an ongoing programme.

Information governance provides strategic direction for information quality drives, sets standards and processes, and ensures goals are achieved.

The six key dimensions of an effective data governance programme include: policy and planning, organisation, standards, processes and methods, monitoring and communication. Without an effective data governance function, information quality efforts are unlikely to succeed.

Given the dependence that manufacturers have on co-operation across the functional areas that comprise the value chain, designing and implementing a sound governance model is not an easy task - but it is absolutely a prerequisite to successful information management.

Companies are looking to BI solutions as a means to help address regulatory compliance issues on an enterprise basis.

A comprehensive regulatory compliance programme must address deeper business and technology issues, including information quality and data integration problems.

A big opportunity exists to leverage investments in BI solutions to support regulatory compliance programmes. Regulatory compliance has become a driving force behind data warehouse and BI programmes in many industries.

Organisations want enterprise-level BI because they need an integrated view of data from many disparate sources, largely as a result of the growing demands for regulatory compliance, increasing merger/acquisition activities, and a desire to monitor and analyse performance at enterprise level.

But, enterprise-level BI requires significant process and organisational changes, and a BI architecture must reflect the input of both business and IT. Again, an iterative approach is key.

Improving customer satisfaction is a key objective for most companies. However, it remains an elusive aim, because most systems simply don't support a single customer view.

When customer data is stored in disparate systems as well as various business applications, users are forced to search for and extract data from all of them. Worse still, the accuracy of the customer view depends on the quality of the data.

What is required is an information management strategy that places key data in one central repository from which all other systems draw. Doing so can create a single valid source for this critical business data. Within such a strategy, customer and business-focused BI tools can help reap the rewards of a single view of the customer.

BI is lessening business risk by enabling organisations to exploit sources of information and knowledge, inside and outside, to forecast outcomes and assess options for advantage.

It ensures both positive and negative risk factors - opportunities as well as threats - are incorporated in strategic planning.

But, this means looking ahead, as well as back, to identify risk scenarios, indicators and both threats and opportunities.

With the increasing commoditisation of BI product suites comes consolidation. That means most BI companies will be offering very similar systems.

It therefore makes good sense for companies to view any new BI purchases in the context of a framework.

The spate of acquisitions in the BI world has resulted in BI players latching onto performance management as a way to differentiate themselves and add value for customers.

BPM gives organisations the ability to discern patterns or trends from organised information, with continuous and real-time reviews helping to identify and eliminate problems before they grow.

BPM's forecasting abilities help a company take corrective action in time to meet earnings projections. Forecasting is characterised by a high degree of predictability, which is put into good use to answer what-if scenarios. BPM is useful in risk analysis and predicting outcomes of merger and acquisition scenarios and coming up with a plan to overcome potential problems.

It provides key performance indicators (KPIs) that help companies monitor efficiency of projects and employees against operational targets. However, it depends on KPI-related data that is consistent and correct, and the timeous availability of KPI-related data, all of which is related to the points discussed elsewhere in this article.

There are indications that extraction, transforming and loading (ETL), the movement of data from source systems into the data warehouse, is becoming obsolete. This is a result of extract-load-transform (ELT), the third generation of ETL. ELT eliminates the need for dedicated ETL servers, by talking full advantage of the power of the relational database management system to perform compute-intensive transformations.

With the ELT approach, data is transformed on the target after being loaded. Indeed, if the target database is powerful enough, it can be used to perform all transformations and optimise both performance and investment.

A business rules-driven approach enables developers to describe end-to-end transformation and data quality rules in terms that are the most intuitive and easiest to maintain. The result is better developer productivity, better performance and lower costs.