Wednesday, December 2, 2015

The Fallacy of The Canonical Dashboard(s)

I've once again come across an article promulgating the conventional wisdom that runs along the lines of: "Important information about the Dashboard (or two, or three) your business needs."
It's here: Why every business needs two dashboards for clear flying, and contains this passage:

The two dashboards every business needs

"But it actually isn’t enough to have just one dashboard; I believe every business needs two dashboards: strategic and operational. Like the cockpit instruments in a fighter jet, they allow the executive to know exactly where he or she is at any given time and focus on getting to the destination in one piece."

Putting aside the unfortunate, and by now antiquated, fighter jet cockpit metaphor, the article recognizes that one dashboard isn't enough. But it continues to promote the idea that there is a small set (in this case: two) dashboards that, if carefully considered, can provide the information decision makers need to run their business.

This is an anachronistic view of the world of business data analysis that doesn't recognize developments of the past decade that have moved beyond its limitations.

In the real world, any small set of canonical dashboards is limited in the information they can convey, and don't extend more than a step or two towards the horizon of useful information.

The idea that there's a limited view of one's information space that's adequate for monitoring and decision-making is rooted in historical factors. Briefly: because it took very substantial amounts of time, energy, money, and other resources required to create information delivery artifacts, e.g. dashboards, people became conditioned to the idea that there was a limited view that, once identified, designed, built, and delivered, would be adequate for their information needs. This was always an artificial limitation, an unfortunate (and in reality unnecessary) consequence of and concession to the deficiencies of the business data management and analysis environment.

The past decade has seen the emergence of better, faster, low-friction, tools, technologies, and practices that dramatically narrow the gaps between data and the people who need to understand it.

The past five years has seen the increasing awareness of these tools, particularly with Tableau's recognition by Gartner, Forrester, TDWI, and related media and general audience channels.

The implications of the new opportunities have, as in all paradigm shifts, been slower to bubble to the surface, but they're starting to become part of the discourse, even as the traditional message that there's a canonical set of dashboards that's sufficient for running a business persists.

The modern reality is that it's possible to discover and deliver data-based information on an ongoing basis, including but not limited to a small set of pre-identified KPIs in one or two dashboards. There's a very small distance between dynamic data discovery and the composition of relevant analyses into dashboards—this is a fundamental departure from the traditional BI world, and marks a qualitative shift in how effective business data analysis can be pursued. It's now possible to provide the information people need to make decisions from the relevant data as they need it, even if it's not previously been formalized in pre-constructed forms: dashboards, scorecards, etc.

Organizations that recognize that they're no longer constrained by the traditional limitations can take advantage of the new opportunities and dramatically improve their data-based decision making abilities. One of the first steps is recognizing that they can access, analyze, and understand their data as needed, rather than speculating about future information needs and spending time, energy, and effort tackling technical implementation efforts for potential payoff. As they absorb this concept, people recognize that they no longer need be shackled to one, two, or some small number of discrete dashboards.