Dashboard design

Two years ago I worked with a Gartner VP and Research Fellow, Ken McGee, on a book he wrote for Harvard Business School Press: Head’s Up, which I have blogged about in the past, but which I feel is essential reading for any analyst or manager concerned with the concept of dashboards for the presentation and reporting of quantifiable performance measures.

I won’t delve into the theory of dashboard design, but McGee offers some salient points for anyone attempting to reduce the information flood generated by an organization into the essential key points that a senior executive needs, as soon as possible, to anticipate surprises or inflection points and react. The purpose of a dashboard, as McGee emphasizes, is not to predict the future, but to predict the present, giving insights to actions occurring in as close to real time as possible. The sooner management is aware of an opportunity or threat, the sooner they can react, and the wider the window of reaction time, the better the chance of seizing the opportunity or thwarting the problem.
One blog, Dashboard Spy, is a wonderful collection of screenshots of working dashboards and offers some great design tips. I strongly recommend it. It too emphasizes McGee’s point that less is more, and that there is not canonical approach to dashboard design that can be templated from one organization to another. The issue is not having enough information, it’s finding the few essential indicators and presenting them in a format compelling enough to drive action. Let me repeat the last point: drive action. Analysis for the sake of analysis without a channel for action or correction is academic information only. The creator and collector of the dashboard and the receipient must constantly be tuning the dashboard to insure it remains a valid tool. Let it rot, unopened in an inbox, and the dashboard is useless.

Author: David Churbuck

Cape Codder with an itch to write

0 thoughts on “Dashboard design”

This site uses Akismet to reduce spam. Learn how your comment data is processed.