How Dashboards Have Grown Up
Visual reports are no longer sufficient. Today’s dashboards must deal with big data and short response times.
By Bob Potter, Vice President and General Manager, Rocket Software
Ten years ago, a dashboard was a visual report that was distinguished by its charts and graphs, but it more or less gave you the same information found in a tabular report. These historical dashboards also collected the data and ran just like the venerable reports did, with complex batch jobs created by an IT developer.
It wasn't so long ago when IT departments "owned" the data and dictated how business users visualized that data -- and how long business users had to wait to get access to the locked-down systems that automated customer management and accounting. Queries were always pre-defined, and business users had to know in advance what they needed to see and how much “drill-down to detail” was necessary.
The business user was always looking at what had already happened, which was not the best way to gain insight into what was going to happen and how the business needed to adapt in order to be agile for changing business conditions. Think of it as driving a car while only looking in your rearview mirror. Users would gather in a conference room and argue about the authenticity of the data, and "collaboration" was defined as "who would concede first on the analytical perspective."
Flash forward to 2014. Dashboards are now highly interactive and business intelligence has evolved into visual analysis, data discovery, and self-service BI. Sure, we still run tabular reports the way we did a decade ago, but business users are now empowered to explore data without really knowing what they are looking for ahead of time. “Data scientist,” a recently glamourized job title, describes someone who can provide visual analysis and data discovery without knowledge a priori, but data scientists are the exception, not the rule.
Data discovery for most business analysts goes something like this: “I think we have reimbursable hospital bills that can be submitted to the state and/or federal government instead of being written off as an uncollectable receivable.” This is not a fictitious use case, but a real-world use case from one of our customers!
This isn't just a hypothetical construct -- it's playing out in the real world every day. The ability for the billing department of a major U.S. hospital to comb through thousands of unpaid bills over a three-year period and find anomalies and characteristics in the data that would allow them to submit a claim for reimbursement was made possible by building a visual application that accessed huge amounts of DB2-managed data on the mainframe -- and this resulted in the discovery of hidden treasure. In fact, there was so much treasure that the reimbursement funds they collected paid the hospital’s entire IT budget for a whole year. Business users, in conjunction with the IT department, built the application in just two weeks. Imagine the ROI from that one dashboard application!
Another key area that wasn't even a blip on anyone's radar 10 years ago is mobility. After all, there was no real point to remotely accessing complex data on a flip phone. Then smartphones and tablets came along, and with them came the brave new world of bring your own device (BYOD), where users prefer to access their data in a visually compelling format and from any device, at any time, and from any location. That's why today's BI tools need to have mobility baked-in. It's no longer a luxury or nice-to-have feature. It's an imperative.
Today’s dashboards need rich, visual interfaces that are intuitive and easy to use, both when developing the applications and when connecting to the source data. Performance must be snappy, even when processing very large data sets containing millions of rows of data.
Users want to search and collaborate with co-workers. I was recently in a meeting with the executive team of my company and we decided that we shouldn’t look at a snapshot of the previous day’s Salesforce.com data but rather analyze current data. Our VP of sales hit the “Refresh” button and we all sat around watching spinning wheels for several minutes. The dashboards never did update. Modern data visualization tools should fetch data several times a day and when you request an update, the dashboard should update in a second or two.
A data refresh longer than five minutes (and the charts and graphs taking more than five seconds to update) is an unacceptable latency. Business users today have little patience when it comes to exploring data. They need actionable information at the right time -- no matter where they are located.
Call it whatever you like: data discovery, self-service BI, visual analytics, or data visualization, those are all good terms. Business users just call it common sense when it comes to getting their jobs done.
Bob Potter is vice president and general manager of Rocket Software's business intelligence/analytics business unit. He has spent 33 years in the software industry with start-ups, mid-size and large public companies with a focus on BI and data analytics.