Business intelligence dashboards are frequently bashed. I slammed them back in 2006 and 2007. Mark Smith dropped the hammer last August. EIS, the most dashboard-like pre-1990s analytic technology, was also the most reviled. There are reasons for this disdain, but even so dashboards shouldn’t be dismissed entirely.
In essence, I’d say:
- Dashboards are overrated and oversold.
- They are useful even so.
- Their usefulness is ebbing as technology advances.
- The metaphor of an automobile dashboard — or airplane cockpit — was always imperfect. A car’s dashboard shows you ALL information that can be surfaced, and also holds pretty much all your controls. But a BI dashboard is just a selective, surface view.
- BI dashboards have two basic uses — giving a snapshot view of what’s going on, and serving as a portal for further BI activity.
- The snapshot view can work well in human real-time scenarios such as stock quotes or network monitoring. Otherwise, the benefits of a snapshot are less clear.
- The portal use of a dashboard is conceptually solid, and I referred to it favorably in a recent BI taxonomy post. However, numerous user interface trends point toward single-chart/single-report BI, including:
- The smaller screens of mobile devices.
- The sophisticated drilldown of a Tableau or QlikView.
- Search-based BI navigation.
- Collaborative BI.
- Integration between BI and traditional packaged applications.
I’d further note that:
- Marketing departments are strong dashboard adopters — but they’re also strong adopters of most other analytic technologies, including the ones I cited as reasons to get away from dashboards.
- Irrespective of their later usefulness in production, dashboards demo really well.
And so my tentative view of dashboards’ future boils down to:
- I imagine that human real-time dashboards will be with us for a long time; even with great alerting technology, you’d want great dashboards as well for rapid information intake.
- I also think we’ll long have BI portals — and the less sophisticated or experienced a user is, the more useful a dashboard-like display could be in showing him which information options are available.
- Beyond those two use cases, I don’t know why we should expect that dashboard technology will ever be “fixed”.