Friendly suggestions to help revitalize and maximize your company's BI strategy.
Posted Oct 2, 2007
Many organizations have adopted business intelligence (BI) tools in quantity, buying thousands of user licenses to deploy up to 50,000 seats of some of the leading BI tools. There's a growing body of evidence that many of these licenses remain unused, and the majority fall short of meeting expectations. Often the main benefit from deployment is merely a secure method of distributing standard reports.
The following list will help you dust-off and maximize your BI investment and provide suggestions for improving overall results.
1. Don't run your BI initiative in a vacuum
Many business intelligence projects focus too heavily on role-based distribution of data; the more senior the employee, the more data they can access. Remove this paradigm, providing employees access to all the information they require unless there is a compelling reason not to.
2. Determine the action you want people to take
Research indicates that most BI reports -- because they look back at history after the event -- are used to justify decisions that have already been made. Avoid this pitfall by thinking through what decisions you want people to make and when. This will help you to decide what information needs to be delivered in what timeframe.
3. Don't blindly recreate what's already there
Just because a report exists today, doesn't mean that you should blindly recreate it when you upgrade or migrate to a new tool or dashboard. Very often the business will use only a tiny fraction of the reports. For those reports that are still needed, consider the information that is really required, and how best to present it given the capabilities of your new tool.
4. Identify which reports aren't used (most) and turn them off
Some BI teams experiment by turning reports off to see if people notice. Try it -- you'll be surprised just how many reports you don't need in your organization!
5. Design from the business back to technology
Most BI tools make it easy to set up the end-user environment by translating the database dictionary into something more user-friendly. As a result, IT staff often build BI systems by simply renaming the structures that exist in the database. This is guaranteed to make the eventual system the business sees too hard to use. Start with the language the business uses, and work back to the database, not the other way around.
6. Teach people about your data
One of the main reasons BI is too hard to use is not the technology tools, but the data itself. The single-most effective way to get people to use a BI system for themselves -- as opposed to asking someone else to run an ad hoc query or analysis for them -- is to train them on the data. Without training, they haven't a hope of being able to generate their own reports and analyses safely.
7. Less is more: Aim for fewer, more relevant, and timelier reports
Large BI vendors recommend that you deploy BI tools onto every desktop and to produce thousands of reports. In fact, some data warehouses produce over 100,000 different reports. Don't be surprised when users complain that the information they get isn't that relevant or timely! The goal of every BI project should be to reduce the number of reports, and to make the content and timing much more relevant.
8. Think about timing
Delivering reports at the end of the day will not help you make strategic decisions at noon. Consider a 2 p.m. cut-off deadline for merchandise reorders. Delivering a report at the end of the day means that you are forcing the business to use guesswork at 2 p.m. rather than optimizing the process with the right information. "Right-time BI" advocates claim that delivering the information just in time is sufficient. I would argue that you should always aim to weed latency out of your systems. After all, the sooner you get the information to the business, the more reaction time they have. This type of operational BI problem should be addressed with a continuously updated system, not one reliant on batch updates.
9. Train people how to analyze data
Most business people are not analysts, yet some basic analysis techniques go a long way in empowering them with their new BI tool. Most organizations don't do this, though the Six Sigma programs run by some companies -- such as GE -- call for many more people to be trained in this way. Their use of information in decision-making changes dramatically after this type of training.
10. Build BI into daily processes
When you fly across the Atlantic on an aircraft, the pilot doesn't actually fly the plane, but relies on an autopilot. He looks at the dials in the cockpit, monitors warning lights, but ultimately lets the autopilot do most of the flying. The autopilot doesn't get tired, but optimizes the mundane and flies the aircraft more efficiently than any pilot could. This has strong parallels with business operations. By building BI into the processes themselves, business runs more smoothly and efficiently. This gives you time to worry about the exceptions and to oversee the process as a whole. In a recent research project, we found that 90 percent of managers believe that there would be benefits to their business if they could embed greater intelligence into daily operational processes.
About the Author
Charles Nicholls, founder and CEO of SeeWhy Software (www.seewhy.com), is a Business Intelligence visionary. He incorporated SeeWhy in 2003 to create a new generation of Business Intelligence to revolutionize the way organizations analyze and use data.
MicroStrategy 9, the business intelligence company's first major release in nearly four years, boasts enhanced features and an increase in both scalability and flexibility.
Convert up to 50 percent of abandoners with real-time follow-up.