Traditional BI has failed

Business intelligence (BI) has failed because it hasn't fulfilled the original promises it made.

This is according to Davide Hanan, MD of QlikView SA, who says many organisations need a complete overhaul of their ageing BI systems that have failed to provide business benefit.

“Traditional BI has made the life of IT easier, but this is certainly not the same case in terms of the user. This is despite the fact that BI was originally built in mind to serve the business user.”

According to an ITWeb survey, only 19% of respondents at the 2010 BI conference, held earlier this year, viewed their BI solutions as a success.

Hanan explains that traditional BI reporting has resulted in information overflow and data complexity. “Around 10 variables can result in 362 8800 permutations or reports. The issue of information overflow is not a function of the amount of data to be analysed, it's a function of the tool being used to analyse that data.”

He says people should be asking for more information not less, but to work with more information, organisations need the right tools to simplify data and access information to answer business queries.

“BI reports and Olap cubes were born in the late 80s, which, during the time, were restricted to hardware limitations,” explains Hanan. “Yet, in 2010, many BI vendors are still selling the same technology developed in the 1980s and these are still being used by companies today.

“Twenty years ago, it was not easy to get data from multiple systems, that was why a data warehouse was formed to bring islands of information together. However, we no longer have these limitations. The constraints we had in CPU power and memory is now gone.”

According to Hanan, CPU performance has increased 25 times in the past five years alone, and memory has become a lot cheaper.

“The only reason why we still have Olap cubes is because of the huge costs that companies would need to invest in to replace their legacy systems. In addition, the knowledge and skills within an organisation prevents them from doing so and old technology protects jobs.”

He adds that the data warehouse itself won't go away in the short term because enterprises are still dealing with disparate legacy systems.

“Data warehouses should then be deployed if needed; not as a default. Data warehouses are used to link data in disparate systems and mapping different data representing similar information in one place.”

The future of BI lies in in-memory analytics. Hanan cautions that BI vendors will need to do more than rolling out just in-memory BI. He says they need to go further and change the underlying architecture and the speed of in-memory analytics needs to be combined with the ability of users to do associative thinking.

QlikView recently reported strong growth, especially after the listing of its parent company QlikTech on the Nasdaq Stock Market on 15 July. According to Hanan, the IPO was well received in the industry and the share price increased by more than 90% in less than two months.

“We have continued to grow at an average of 15 new customers per month this year, and recently passed the 500th customer milestone,” says Hanan.

At the time of publication, QlickTech's share price was $20.9, which Hanan says more than doubled from two months ago.