Analytical business intelligence (BI) software capable of not just reporting on corporate data but also automatically analysing it has been around in various forms for almost as long as the query and reporting tools that have become largely synonymous with the BI sector. But it is only in recent years that several factors have combined to make analytics a priority for many IT departments.

Tom Davenport, an expert on BI systems and author of the recent Harvard Business Review article Competing on Analytics, argues that three developments have made analytical BI systems increasingly viable. These are the development of 64bit computing systems capable of analysing larger data sets; improved organisation of data in business applications; and the emergence of a new generation of executives who make decisions based on metrics.

These three factors have made analytical BI technically and culturally feasible but it is the realisation that they can deliver major competitive advantages that is driving the current surge in demand.

Tesco reportedly saw a 20-fold increase in responses to sales promotions when it used a BI system to analyse customers' loyalty card data. Meanwhile, Barclaycard is using analytics to avoid customers likely to default on their debts. Such examples have helped many firms to see that better analysis of data can directly improve the bottom line. There is a widespread realisation that such analytical BI projects really do deliver steep improvements in business metrics, says Davenport.

The ability of analytical BI systems to bolster firms' competitiveness is clear.
However, IT directors face many challenges before they can successfully implement BI systems and realise the commercial benefits promised by vendors, such as BI specialists Business Objects, SAS, Cognos and Hyperion, and larger providers, such as Oracle and SAP, which have embedded analytics into their business applications.

The first challenge is to select the right analytical BI system. Phil Wood, product marketing manager at Business Objects, believes analytics has become an umbrella term covering many different technologies – so firms must be careful to choose the right system for their environment. "It is a case of using the right tool for the right job," says Wood. "If you want to find out which is your most overstocked and understocked store you only need simple reporting. But if you want to accurately forecast future demand you are going to need high-end data-mining capabilities."

Those firms that deploy high-end analytics will then have to undertake a lot of groundwork to ensure accurate results from their new systems. According to Russ Cobb, senior product marketing director at SAS, such deployments may prove a waste of time if firms do not carry out the data integration work needed to ensure their data is clean. "If you apply analytics to bad data you will get bad results," he explains. "You have to ensure metadata management and central repositories are up to scratch." 

According to a recent SAS survey of 1,100 executives, many firms are failing at this stage. Less than half said their firms had data cleansing and rationalisation processes. As a result the success of analytical BI deployments is often undermined.

"A lot of firms don't have the basics in place in terms of data quality and integration to underpin predictive analytics effectively," agrees Julian Highley, product manager at Business Objects. "[Analyst firm] Gartner has said that any company that thinks it doesn't have a data quality issue is in denial and it's right."

Once data integration and cleansing technologies are in place, IT directors need to ensure they have the right people in place to install and operate the technology. As Christina McKeon, product marketing manager for BI at SAS, explains, "There are no layman's terms to explain the algorithms [used in analytics systems]. To make them work effectively you need some very highly skilled staff."

But as well as investing in skilled statisticians, IT chiefs must ensure the results of the statisticians' work reaches end-users, says Mark Morton, senior product marketing manager at Cognos.