DIY BI: A guide to self-service business intelligence implementation
A comprehensive collection of articles, videos and more, hand-picked by our editors
Despite all the hype about self-service business intelligence, empowering business users to create their own analytical reports and queries remains a challenging and elusive goal for many BI managers, according to Wayne Eckerson, director of TechTarget Inc.'s BI Leadership Research unit. Last year, Eckerson authored a report on the use of self-service BI tools, based on a survey of 234 BI and IT professionals, business users and consultants that he conducted in July 2012. In the report, Eckerson wrote that only 36% of the survey respondents gave their BI self-service programs a grade of "good" or "excellent." One of the reasons, he said, is that implementing self-service BI capabilities "is more complex than it looks."
In an interview with SearchBusinessAnalytics, Eckerson expanded on the complexities of managing self-service BI initiatives and said that segmenting end users into separate groups and supporting different levels of functionality based on their varying needs can go a long way toward ensuring an effective implementation. In other words, one size fits all is not the path to self-service success. Excerpts from the interview follow:
Your research shows that many companies aren't satisfied with their self-service business intelligence deployments. What's the underlying reason for their dissatisfaction?
Wayne Eckerson: A lot of companies find self-service BI harder to do than they thought. I've heard from BI professionals that self-service BI takes a lot of handholding -- well, that's counterintuitive to the notion of self-service. Most of the tools are not easy to use, even for power users who get paid to write reports. You have to empower users, [but] you have to train them and continue to support them because it takes a long time to become proficient with these tools.
You have to empower users, [but] you have to train them and continue to support them because it takes a long time to become proficient with these tools.
Why isn't a one-size-fits-all approach the right strategy for self-service BI adoption?
Eckerson: Ten years ago, there was a move among IT managers to standardize on tools and minimize the number of suppliers in BI [environments], the goal being to have one tool that meets everyone's needs. We realized that doesn't work. You need a number of tools for data mining, ad hoc analysis, dashboards. The question with self-service BI is, who is that tool geared for -- report users or report authors? If you give a report user a tool designed for report authoring, they're not going to use it because it's too complex.
How do BI data requirements differ for those two groups of users?
Eckerson: Report users use information to do their jobs, and the kind of self-service they want is more interactivity with the information in a report. The interactivity they want occurs in gradients; think of it like a hierarchy. Some want to view a report, print it off and circle a few things; some want to navigate to the next level of detail; some want to use the data but look at it in a different way; some want to add new data to a report; and some want to do complex modeling inside the report.
Report authors want to do more sophisticated things to create new reports, from sourcing new data to potentially doing coding. Different people are comfortable with different hierarchy levels for how they want to use information, and the best BI tools expose information on demand so you don't overwhelm lower-level users. It could even be the same set of tools, but different functionality gets exposed depending on the user.
How does this concept of hierarchies affect deployment strategies for self-service BI tools?
Eckerson: What it does is help [BI managers] really understand users and the roles they play and the information they need in each of those roles. Some users play multiple roles, and each might have different information requirements. The idea is to deploy a tool and configure it using role-based access so users can see the data functionality they need and want but no more.
What role can "superusers" play, particularly in helping to accelerate adoption of self-service BI capabilities?
More on managing self-service business intelligence projects
Learn about the self-service BI strategy adopted by mortgage lender Quicken Loans
Read how using self-service applications can reduce BI workloads for IT staffs
Get advice on governing self-service BI usage to avoid data chaos
Eckerson: Superusers are business users in each department who gravitate toward a BI tool when it's deployed and quickly become the go-to people in their department to create ad hoc views of data. Most casual users today don't have the time or patience to learn how to use these tools, so when they want to look at data in a slightly different way, instead of calling IT and waiting a month for the report, they can call their buddy and he'll do it for them.
Superusers are one of the keys to a successful BI program. They're the eyes and ears of the BI team in the trenches. Ideally, they function as virtual members of a BI competency center, playing a very valuable role in helping implement BI in a systematic, nonchaotic way.
How do you know when a self-service business intelligence program is a success?
Eckerson: Your backlog [of query requests] goes down, users aren't resorting to spreadmarts to get data and are leveraging corporate-issued BI tools, and there is no report chaos. If you have those three things, you've succeeded.
Beth Stackpole is a freelance writer who has been covering the intersection of technology and business for more than 25 years.