During a 2009 Gartner Inc. business intelligence (BI) conference, there was talk of a new class of data discovery tools that would allow a broad range of users to search for information and create their own dashboards and reports. Today, many call this approach self-service BI, and it could be the key to finally unlocking BI's ROI potential. It won't happen, however, without an expert hand at the controls.
You can't put a BI tool on everyone's desktop hoping that they'll 'teach themselves to fish.'
Take the case of investment firm RBC Wealth Management USA, a Minneapolis-based division of RBC Capital Markets LLC with 6,000 employees. Since it introduced Tableau Software Inc.'s Web-based Tableau Server data analytics tool, the number of RBC employees making requests to the BI group has gone from 40 to 450. Employees don't have to be experts at BI tools or learn to do complex data queries. Rather, they request the data they want and are provided a Web link to build a customized report or to manipulate the data into their own charts and dashboards with drill-downs.
The uptake of the self-service BI tool has been impressive, said Shawn Spott, vice president of corporate intelligence at RBC Wealth Management. But it doesn't happen without expert guidance and plenty of work on the back end. "You can't put a BI tool on everyone's desktop hoping that they'll 'teach themselves to fish.'"
To keep the BI requests coming at RBC, Shawn's two-person team guides users through 1.25 terabytes of data that employees use to measure individual business lines and overall corporate performance. Staff might see only a Web link, but Spott's team has integrated eight years' worth of back-end data from the mainframe, SQL Servers, a Teradata Corp. data warehouse and an SAS Enterprise BI Server from SAS Institute Inc.
Spott introduced the self-service BI program slowly. He focused first on the company's wealth management consulting group, whose members had become disenchanted with BI after "never having a request met" by the company's existing tools. The consultants wanted to identify the highest-grossing performers among a base of 2,200 brokers with whom the wealth management group's internal and external sales team worked.
"They knew who the highest-grossing brokers were -- say, brokers A, B, and C -- but they had no underlying metrics to show that broker D was also a great prospect. They were desperate for a holistic view of those brokers in the branch offices, the ones they didn't know but should," Spott said.
Other groups saw the success the wealth management group had with the self-service BI tool's ability to identify and reach out to high-performing brokers. Soon, groups such as human resources and compliance were asking for access. "I was surprised by compliance, since they are a pretty closed loop," Spott said.
Comparing self-service BI and traditional BI
When used successfully, self-service BI tools address one of the biggest frustrations users experience with traditional tools: the lag time between a request for a report or dashboard -- or even a simple change to an existing report -- and the response to that request from IT or the BI team.
"It's not just that companies are short-handed on the BI IT side, but it's also the fact that users increasingly want to do more things with data," said James Kobielus, senior analyst at Forrester Research Inc. in Cambridge, Mass. Self-service BI is the "hottest segment of the BI market" among his client base, he said. It "gives users their own tools to pull data from source applications directly, or from an intermediary database that IT sets us for users."
When IT is taken out of the report and dashboard development process, the BI process is more agile, meaning companies can put their insights into action faster, Kobielus said.
More about BI data analytics strategies
Self-service BI not only puts the power in the hands of the user, but also opens up new ways of analyzing data, Kobielus said. Many users have been exposed to visualization, reporting and modeling to a degree in Excel spreadsheets, but they haven't been given easy access to predictive analytics tools and statistical modeling. "These self-service BI tools are embedding predictive modeling and statistical modeling in self-serve, highly user-friendly drag-and-drop user interfaces," he said.
Of course, IT still does a lot of things behind the scenes, as Spott's approach demonstrates -- integrating data, developing templates and showing employees how to use self-service BI tools' drag-and-drop wizard-driven interfaces.
For Spott, the power of self-service BI is not in its GUI -- which he believes is pretty much the same across the available tools -- but in how well it handles metadata to connect to multiple data sources. In his view, that is the biggest challenge for IT groups when they choose a data tool. "What I was looking for was a tool that could make dynamic connections on the back end to multiple data sources across multiple platforms," he said. "That's where the challenge usually comes in: being able to join different data sources simultaneously."
Let us know what you think about the story; email Christina Torode, News Director.