The most common obstacle standing between companies and successful business intelligence (BI) deployments isn’t slow query performance, company politics or even a lack of end-user skills.
It’s bad data, according to a recent survey by the U.K.-based Business Application Research Center, or BARC.
“The most noticeable change is that we nearly always record that the biggest complaint is query speed,” said Barney Finucane, a BARC analyst and lead author of the BI Survey 9. “This year, for the first year, we saw that it was data quality issues.”
BARC has been conducting the survey of BI end users since 2001. This year’s survey included responses from nearly 2,200 end users, or consultants on behalf of end users, most from Europe and North America.
Bad data means big business intelligence problems
Of those, 18% reported poor data quality as the single most common BI problem hindering their deployments. Not far behind in the list of top BI-related problems were slow query performance and company politics, which were each cited by 16% of respondents.
Still, 70% of respondents said their BI projects had met or exceeded business goals. Just 6% said their BI projects hadn’t met any of their objectives.
The survey also tried to identify the most common business benefits end users enjoy. According to the report, "‘soft’ benefits, like faster or more accurate reporting and better business decisions, are much more likely to be achieved than ‘hard’ bottom-line benefits like reduced costs and headcounts.”
The least reported benefit was headcount reduction, good news for BI pros worried that automated analytics technologies could eliminate their jobs.
Fifty-five percent of respondents reported evaluating BI products from multiple vendors before deployment, a slight rise from last year. The most common BI products included in evaluations were Microsoft SQL Server Analysis Services, Microsoft SQL Server Reporting Services, and IBM Cognos Analysis and Reporting.
Among large enterprises, that list also includes SAP BusinessObjects Web Intelligence.
Interestingly, the median volume of data associated with BI projects was just 6.6 gigabytes. While data warehouse vendors in particular have dubbed 2010 the era of “big data,” just 5.2% of respondents reported data volumes equaling or exceeding one terabyte.
“Though there has been a clear increase since 2005, there is no evidence of ‘exploding data volumes,’” Finucane wrote.
Pervasive BI is still lagging in numbers
The survey also indicated that BI has yet to come to the masses, despite BI vendor marketing pitches to the contrary. Only 11% of respondents reported BI deployed to over 50% of employees in their companies. However, 34% of respondents did say they planned to purchase more BI-related licenses in the future.
Enterprise-wide BI deployments in which companies standardize on a single BI product are also few and far between. According to the survey, just four products – SAP Business Warehouse/BEx, Microstrategy, SAP BusinessObjects Web Intelligence, and IBM Cognos Reporting -- had more than 100 users in at least 50% of their deployments.
“I think there is kind of a movement in the industry toward less centralized projects,” Finucane said. “I believe that a few years ago there was a big push toward the idea of an enterprise solution. In recent years, there’s been kind of a backlash.”