Dashboards aren't dead, they're just demoted

People love binaries. Any time a vendor ships a new tool that does some part of the data puzzle better than a dashboard the military-dashboard apparatus comes out of the woodwork with combative articles: “Dashboards are dead!” “Dashboards aren’t dead!” “I love my dashboards!” “New tools are coming for our jobs!”

Look, I get it, as humans we think in patterns and extremes. It’s difficult to accept that sometimes the answer to an “either/or” question is “a bit of both!” Plus, some vendors have, as a marketing tagline, stated that they see dashboards as a thing of the past (Ironically, those vendors also sell dashboard products).

Dashboards and the Peter Principle

The Peter Principle is the idea that people in a corporate hierarchy tend to be promoted to their “level of incompetence” where they are no longer effective enough to be promoted further. It is a satirical idea that has proven to be true in many corporate environments. If a person is competent in a role, they’ll be seen as valuable, they’ll eventually be promoted, and they will continue to be promoted until reaching a level at which they are no longer competent. I’m sure we’ve all met people in our careers who fit this definition.

The Peter Principle applies to dashboards.

They were undervalued when most data analysis was distributed in row-and-column, printed or pixel-perfect reports that did little to communicate anything in the form of actual insight.

Dashboards hit the scene and were immediately valued - offering a faster way to consume information and a way to layer interpretation into a report. They turned data people into heroes delivering interesting, insightful stories. They turned data-driven reporting into something business people could actually understand.

They were promoted. From a productivity tool for analysts to a departmental system to an enterprise-grade platform. Today, most businesses have strategic dashboard products that deliver reporting to everyone in the organization. But as they’ve been promoted, they have found themselves less able to make an impact. Today, many are fed up with dashboards. There are too many of them. They take too long to build and update. People really just want answers.

It makes sense: “improving understanding of certain metrics” was a good job description for dashboards. Today, they’re being asked to “answer everyone’s questions across every possible use case.” It’s too much for them to handle.

The data tool re-org

Let me be clear: I’m not a monster. Let’s not be morbid. I don’t think this is an “Ol’ Yeller” situation.

My proposal: let’s keep them on the team, but right-size their role. Let’s have a re-org, not a funeral. Let’s give dashboards a project they can succeed at.

Dashboards are great for:

• Reporting numbers that have to be defined the same way, over and over.
• Showing progress against a goal.
• Highly constrained and governed analysis where you don’t want people going off the rails.
• Parameterized reporting where teams need to drill down in clearly defined hierarchies or filter paths.

Dashboards are terrible for:

• Ad hoc questions where the developer doesn’t know in advance what question is going to be asked.
• More advanced analytics where it’s important to use data science techniques or experiment with multiple approaches.
• Establishing KPIs for reference in other reports.

Turning dashboards into a team player

Some insights are best delivered in a carefully designed dashboard. Sometimes it’s better to simply model some data and let users ask their own questions. Other times, analysts need the ability to go “off the rails” and explore without models.

All of these tasks are connected - data people build insights, share them in the form of models, apps, and dashboards, and their stakeholders self-serve off the results. There’s always a moment when the stakeholder has something new they need help with, at which point they re-engage their trusted data team, and the cycle starts anew.

Our goal at Hex is to turn dashboards into more of a team player, and our AI analytics platform is designed to support this cycle.

If you’re interested in this new approach to data, let us know. We’d love to chat.