Data visualization tools are certainly eye-catching. The last couple of years have seen the emergence of several software packages that allow users to present data in flashy animations and colorful graphs. But visual impact should not be the only concern. Speakers at the TDWI BI Executive Summit in San Diego said businesses should focus on implementing systems that provide real value and solve real problems.
Taylor Erickson, vice president of IT at Broomfield, Colo., spine surgery supply company Lanx Inc., said visualizations need to compress large amounts of data into small, easily digested pieces. When Erickson first arrived at Lanx in 2011, he helped get a faltering SAP Business Objects implementation back on track. When the system started delivering its first reports, however, he got a call from the head of sales, who said the information was accurate but practically useless. Sales representatives were having serious problems sifting through the 20-page reports to find the pieces of information most relevant to their work.
I think [visualization] is kind of the final stretch of BI, where you have your data but then you're debating it, you're exploring it collaboratively. I think that's the last mile of BI.
founder, BI Scorecard
Erickson ultimately heeded the advice of a colleague who once said that "for sales reps to get it, you have to write it in crayon." Erickson realized that most salespeople are visionaries -- big-picture people who don't like to get caught up crunching too many numbers. He saw that a simplified approach would be necessary.
Erickson decided that implementing a mobile BI platform would be best. The sales reps who would be the primary users of the BI system were typically out of the office on sales calls, so mobility was important. This approach had the added benefit of making the data more easily digestible through straightforward visualizations that could be pushed through to sales reps' mobile devices.
On top of its Business Objects implementation, Lanx implemented a mobile BI platform from San Diego-based Roambi. Simplicity was one of the key factors in the success of the initiative. Erickson wanted it to be as simple to use as the popular smartphone game Angry Birds. He wanted users to be able to view, manipulate and process data quickly without the need for technical expertise. The simplicity, combined with a rapid, three-week rollout allowed Lanx to quickly get value out of its BI systems.
"Those two things allowed us [in IT] to leapfrog and hit this innovation curve and have a more strategic seat in the company," Erickson said.
It's important for mobile BI applications to offer simple and intuitive visualizations, according to Erickson. This is especially true when one considers the limited screen size of mobile devices. But effective visualizations are also important in in-house implementations of BI software.
Data visualizations are about more than just how pretty the information looks, according to Cindi Howson, founder of BI Scorecard. At their best, visualizations help users retain information and make better-informed decisions. Graphic displays can be understood much more quickly than simple data dumps, and well-designed visualizations do this even better, enabling collaborative data discovery.
Read more about data visualization tools
Dashboards and visualizations are natural fit for BI
Google, Tableau help visualization go mainstream
Visualizations key to success of big data projects
"I think it's kind of the final stretch of BI, where you have your data but then you're debating it, you're exploring it collaboratively," Howson said. "I think that's the last mile of BI."
Businesses should exercise caution when choosing a vendor and implementing a system. According to Howson, visual data discovery is a trendy topic that is generating plenty of buzz, and most vendors now say their BI systems have some type of visualization capability. But most organize data into simple graphs, which falls short of the ideal most people have in mind when they think about data visualization tools.
Businesses considering implementing visualization software should consider factors like the visual appeal of the interface, the number of graphs or other elements that can fit on a single page, the number of chart types the software supports and the level of interactivity the charts facilitate, Howson said. IT buyers should also think about whether they want reports to be viewable on mobile devices. Many visualizations are presented in Adobe Flash media types, which many mobile devices do not support.
Howson added that data visualizations tend to work best with in-memory analytics systems, because the speed of in-memory systems allows nontechnical users to experiment without getting frustrated. Users can quickly pull in data from multiple sources, sort it in different ways and apply different visualizations until they find what works best for them.