News Stay informed about the latest enterprise technology news and product updates.

Business intelligence software deployments: Seven fatal flaws to avoid

There are plenty of stories about failed business intelligence software deployment projects. Gartner analysts outlined the problems and solutions at the annual Gartner BI show.

Add to My Yahoo!

Add to Google CHICAGO -- Companies seeking business intelligence (BI) success should proceed with caution.

At the 2006 Gartner BI summit, analysts shared BI implementation anecdotes from companies that have painted themselves into a corner.

"It's the things you don't see that will get you," said Ted Friedman, research vice president at Stamford, Conn.-based Gartner Inc.

In a presentation entitled "The Five Fatal Flaws of Business Intelligence," Friedman and Bill Hostmann, a fellow research vice president, actually outlined seven fatal flaws to avoid. The analysts explained that they had just chosen the most common problems from companies that have "been there, dumb that."

Delusional deployments of business intelligence software

"If we build it, they will come," is a common misperception about BI projects, said Hostmann. Some companies believed that if they built a BI system -- without considering business, user or training requirements -- the value would be so obvious that users would clamor to learn and use the system. Not true, Hostmann said.

"If the number of users is less than the number of years spent on the project, you have a problem," he joked.

For more on business intelligence software deployments

Read coverage of the Gartner business intelligence summit keynote address

Learn how to establish a business intelligence competency center

Find out business intelligence software trends in a 2007 podcast with Gartner

Learn the top seven secrets to successful BI deployments

That is, if a BI project takes four or five years, and only four or five power users are using it, it's a failure. While that's a dramatic example, he said he's seen companies fail to win user adoption for reasons like lack of enterprise-wide cooperation, inconsistent executive sponsorship and implementation issues.

Companies can avoid this fatal flaw by starting with a solid business case for BI, carefully considering requirements and strategically aligning BI with business problems. Companies should design the system to appeal to the "mass" user and then expand the capabilities for the power user, Hostmann said. The RACT test -- will the BI system really be relevant, accurate, consistent and timely to users -- can be helpful, he said.

Cultural calamities

Some people in an enterprise apparently aren't really looking for a single version of the truth. It can be easier for them to work with common assumptions and "dance with numbers" during management meetings, Hostmann said. There are cultural challenges inherent in deploying BI systems, he explained. Don't take them lightly. A true enterprise-wide understanding of the project and how it will positively transform the business is critical to BI success.

"Get the CEO on down to talk about the single version of the truth," Hostmann said.

Data demons

Data quality is a major inhibitor of BI projects, which can cause user distrust and abandonment of the system, Friedman explained. Flawed data can also have dire effects on a business.

"Bad data truly does breed bad decisions," he said.

Potential solutions include implementing tools and techniques like a data quality firewall, which puts controls and measurements on the quality of data going into or out of a data warehouse, he said. It's also critical that companies deploy a sound data governance and stewardship plan in conjunction with BI activities.

Standardizing business intelligence software sellers

Companies that think implementing BI from their existing ERP, CRM or other vendor is a "no brainer" could be in for a surprise, Friedman said.

"BI is not a 'one size fits all' solution," he cautioned.

Companies should evaluate a range of vendor systems to find the BI application that truly fits their needs and business requirements. Skipping this step in favor of an incumbent could be far more expensive in the long run, he said.

Promoting evolution

BI is an activity, not a project, Hostmann explained. Companies need to think of BI as an "evolutionary" process, and should plan and budget accordingly.

"Fifty percent of the requirements are going to change within the first eight months to one year," he said.

Be prepared for change -- and set management expectations accordingly. Managing these inevitable changes requires a flexible infrastructure, he said.

Over outsourcing

Companies that say "we can outsource the whole darn thing," should think again, Hostmann said. Service providers can complement a skills shortage and provide "re-usable methodologies," but they are no replacement for knowledgeable employees.

This sentiment was echoed in a best practices session presented by Henry Groot, director of decision support systems at Trinity Health, headquartered in Novi, Mich. Its BI team tried the outsourcing route, but ended up bringing the project back in house.

"We brought in an outside group to manage the entire development and construction cycle [of our BI initiative]," Groot said in an interview after his presentation. "What I saw -- and I've seen it more than once -- is that we knew more about the business than the consultants did. I don't think healthcare companies and technologists give themselves enough credit for what they know. I use consultants for something I don't know."

Demanding dashboards

The analysts poked some fun at themselves when they said that some managers attend events like Gartner's BI summit and come home with big ideas that their organizations aren't ready for quite yet.

"They say, just give me a dashboard," Hostmann quipped.

The danger is that an impressive graphical dashboard, without the required data integration and architecture behind it, can become just another flawed reporting tool. Incorrectly implemented dashboards that display conflicting numbers and inconsistent data can cause a lack of confidence in BI. That can undermine long-term efforts. The remedy? Strategically planned BI.

Dig Deeper on Business intelligence best practices

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.