BI projects struggle to show value claims survey
More than three-quarters of business intelligence projects take more than a year to produce meaningful results. And nearly two-thirds of companies who have installed business intelligence systems have done such a poor job they get regular complaints from business users.
That's according to a new survey from Kognitio, which found that, in many cases, BI's reputation as a catalyst in IT disasters has been well merited. "I've seen surveys that suggest 60 percent of BI projects fail," said Kognitio's CEO Roger Llewellyn. "This survey validates what we thought. BI is a high risk business. We're trying to remodel how people plan BI and bring down to project level."
According to the survey, 78 percent of respondents said that it took over a year to see meaningful results, often making the data out of date by the time it has been received.
The survey also revealed a staggering gap between the business and IT sides of the business when it came to what was expected from a BI project. According to the survey, 88 percent of enterprises admitted that business departments did not properly define what they want to achieve from an implementation, the assumption being that the project scope would be defined by the IT department.
Llewellyn said there was a fundamental gap between the IT and business departments and that the people who should be bridging it are the business consultants, but who are not succeeding.In fact, Llewellyn believes that the business consultants have a large part to play in the number of BI projects that are failing. "My biggest complaint about the industry, is that there has to be so much consultancy around data warehouses and that's because IT is indelibly linked to consultancies. He said that BI should be about bringing out the data where any question could be asked of it. "To paraphrase a former US president - it's the data, stupid," he said.
He pointed out that Moore's Law - which seems to apply to so many aspects of IT -doesn't seem to apply to data warehousing. You get more and more powerful processors and yet it still seems to take an age to extract meaningful information, he said.
As befits a company that has made its name selling data warehousing as a service (DaaS), Kognitio's Llewellyn thinks that a more flexible approach to BI is needed. "People should take a pragmatic approach to BI," he said.
"You might for example to your developing in a DaaS environment because you've achieved some results in a project, it becomes easier to justify to the board." He said that companies should think small at first and think short time periods. "If you can show value for your investment within 90 days, that immediately lowers the risk."