Dashboard Yea or Dashboard Nay?
Haven’t we been here before (20 years ago), when executive information systems first came to the fore? Yes and no, experts say.
The dashboards of today differ fundamentally from the relatively static dashboards of yore. According to some proponents, in fact, today’s dashboard is the killer app that could finally take BI mainstream.
"Dashboards resonate with users and are the tool that will finally make BI pervasive," says Wayne Eckerson, director of TDWI Research and author of Performance Dashboards: Measuring, Monitoring, and Managing Your Business
(Wiley, 2005.) "Dashboards conform to the way a majority of users work with information and don’t force them to conform to the way BI tools work."
Of course, not all industry watchers say that dashboards have fully arrived. Some folks say that predictions of dashboard ubiquity—much less of dashboard inevitability—are wildly optimistic. After all, they argue, simplicity of the dashboard interface frequently (and misleadingly) masks the complexity of the underlying infrastructure. To the degree that dashboards are both usable and actionable, they’ve first got to lick infrastructure connectivity problems—such as sourcing fresh (or mostly fresh) data from transactional systems—that have bedeviled the industry for years.
There’s also the déjà vu factor, too. Haven’t we been here before, 20 years ago, when executive information systems (EIS) first came to the fore?
Yes and no, experts say. Perhaps the most suitable comportment to dashboards and dashboarding is cautious optimism of a kind. Yes, the dashboard is a powerful interface and, yes, it’s surprisingly appropriate for a wide range of users. And, yes, when yoked to Web 2.0 technologies it provides a powerful means to quickly and relatively painlessly expose new applications—and, even, to empower users to build applications of their own.
But the dashboard isn’t a radical (or even transformative) recasting of the information delivery status quo. Let’s call it a revitalized, more usable, and much more interactive spin on the EISes of old.
Put Mark Madsen, a principal with consultancy Third Nature Inc., in the cautiously optimistic camp. Madsen isn’t a dashboard triumphalist—not in any sense of the word—but he does admit being impressed by the flexibility, usability, and rapid application development-ability of the new Web 2.0 tools.
"I won the mashup competition at Mashup 4 [held in late July] with an app that pulled live Salesforce data, looked up competitors, competitor news, current news for the selected company, found restaurants within 10 miles of the office of the SF.com prospect and checked the weather—all interactive and driven directly off SF.com," he comments. "If I had another eight hours, the UI would be nice and completely automated, as well as doing call generation and logging and feed results back into SF.com."
That wasn’t the most intriguing aspect of the project, either, says Madsen, who has authored or co-authored several books, including Clickstream Data Warehousing. "The [total] time to develop [was]: [approximately] 8 hours, including all data sourcing from a half-dozen external sources. Last time I wrote a program: 18 months ago. Tool cost: $0."
Madsen used a number of free Web 2.0 tools to build his sample project. Based on this experience, he says he’s impressed by what’s out there. "I think the Web 2.0 tools are still very early, but they do the same things all these dash displays do, and do them with less overhead and more flexibility," he comments. "The thing they are missing is metadata and drilling into details, but with Webified everything I can easily go from a graph widget to a report in a crusty old web 0.5 BI tool, as well as simply accessing remote data."
This enables a new kind of RAD design. "When things can be built this fast, throwaway code is possible, and new and more sophisticated customized contextual apps can be done. The IBM guys are calling them ‘situational apps’. Makes me think the BI tool thrust with universal interfaces will continue to miss unless they start working on re-architecting their products," he argues.
Madsen is speaking about Web 2.0-boards, which—in terms of sheer what-you-see-is-what-you-get-ability (WYSIWYG)—are in a whole different class from the packaged dashboards marketed by most BI and PM players. Not that packaged dashboards aren’t being (or haven’t been) Web 2.0-ified, of course. Most major players—including Business Objects SA, Cognos Inc., MicroStrategy Inc., and Oracle Corp. (which purchased dashboard champ Hyperion Solutions Corp. earlier this year), among others—deliver Web 2.0 features sets in their own dashboard offerings. These include support for both Asynchronous Java over XML (AJaX) and Adobe Flash, along with the requisite support for Web services standards like the Simple Object Access Protocol (SOAP), eXtensible Access Control Markup Language (XACML), and others.
At this point, anyway, the current crop of dashboard tools aren’t yet as turnkey as the Web 2.0 model that Madsen—himself a product of Carnegie Mellon University’s prestigious graduate program in computer science—used to build his award-winning Mashup proof-of-concept.
"The tools are better and easier to set up and manage. However, I would be cautious about claiming that usability is not a problem," says industry veteran Rajeev Rawat, founder and CEO of consultancy BI Results LLC.
Like Rawat, Tony Politano, author, lecturer, and partner with BI consultancy BusinessEdge Solutions Inc., says he likes what he sees in the dashboards of today. But Politano, too, cautions against dashboard triumphalism—or fait accompli-ism—of any kind.
"From a technology usability perspective, dashboards are in an enviable position. They combine a very approachable UI along with focused metrics and measures. This is a good first step in licking the usability challenge," he comments. The rub, Politano says, is more or less the same one which Rawat highlights: the gap between implementation and execution. "If organizations build on this usability win and drive it into their BI, it is a great one-two punch," he points out. "The challenge … is that even when they have outstanding usability on the dashboard, as soon as the user wants more information and clicks out to the BI tool, it is to the same techno-centric screens. We need to bridge this gap to achieve the usability bridging from dashboards to BI."
BI tools guru Cindo Howson comments: "I wouldn’t say they are the killer interface—remember the [dashboard’s] ‘killer interface’ looks a lot like the EIS’s of the 80s. They are an important interface though, and the way in which they now get built—by users themselves—and the frequency at which they get updated [i.e., real time in some cases]—makes them much more appealing now than in the past."
All the same, Howson, principal of biscorecard.com, says she doubts whether dashboards are appropriate—or necessary—for all users at all levels of an organization. "Not [for] all users—only [for] users who need to view multiple sets of information simultaneously, which often means managers. Some people who are only responsible for one thing may only need a single report, or a widget on a portal, or an alert on a Mobile device," she comments.
Dashboard champion Eckerson applauds the intuitiveness of dashboards, which he says makes them suitable for users of all classes—from rank-and-file knowledge workers to managers to executive decision-makers.
"To me, dashboards conform to the way casual users prefer to work with info. They assume you know your business processes and what you need to monitor on a regular basis," he indicates. Of course, the dashboard isn’t an end-all-be-all unto itself. Eckerson says that another technology has emerged as a complement to the increasingly ubiquitous dashboard.
"BI search will become the ad hoc complement to dashboards where you don’t know what you are looking for in advance," he concludes.
Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.